agora inbox for [email protected]help / color / mirror / Atom feed
[PATCH v6] README for 64bit xid 265+ messages / 2 participants [nested] [flat]
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v6] README for 64bit xid @ 2022-01-10 19:20 Pavel Borisov <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Pavel Borisov @ 2022-01-10 19:20 UTC (permalink / raw) Authors: - Pavel Borisov <[email protected]> - Maxim Orlov <[email protected]> - Yura Sokolov <[email protected]> <[email protected]> --- src/backend/access/heap/README.XID64 | 128 +++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) create mode 100644 src/backend/access/heap/README.XID64 diff --git a/src/backend/access/heap/README.XID64 b/src/backend/access/heap/README.XID64 new file mode 100644 index 00000000000..457ba9b9ef5 --- /dev/null +++ b/src/backend/access/heap/README.XID64 @@ -0,0 +1,128 @@ +src/backend/access/heap/README.XID64 + +64-bit Transaction ID's (XID) +============================= + +A limited number (N = 2^32) of XID's required to do vacuum freeze to prevent +wraparound every N/2 transactions. This causes performance degradation due +to the need to exclusively lock tables while being vacuumed. In each +wraparound cycle, SLRU buffers are also being cut. + +With 64-bit XID's wraparound is effectively postponed to a very distant +future. Even in highly loaded systems that had 2^32 transactions per day +it will take huge 2^31 days before the first enforced "vacuum to prevent +wraparound"). Buffers cutting and routine vacuum are not enforced, and DBA +can plan them independently at the time with the least system load and least +critical for database performance. Also, it can be done less frequently +(several times a year vs every several days) on systems with transaction rates +similar to those mentioned above. + +On-disk tuple and page format +----------------------------- + +On-disk tuple format remains unchanged. 32-bit t_xmin and t_xmax store the +lower parts of 64-bit XMIN and XMAX values. Each heap page has additional +64-bit pd_xid_base and pd_multi_base which are common for all tuples on a page. +They are placed into a pd_special area - 16 bytes in the end of a heap page. +Actual XMIN/XMAX for a tuple are calculated upon reading a tuple from a page +as follows: + +XMIN = t_xmin + pd_xid_base. (1) +XMAX = t_xmax + pd_xid_base/pd_multi_base. (2) + +"Double XMAX" page format +--------------------------------- + +At first read of a heap page after pg_upgrade from 32-bit XID PostgreSQL +version pd_special area with a size of 16 bytes should be added to a page. +Though a page may not have space for this. Then it can be converted to a +temporary format called "double XMAX". + +All tuples after pg-upgrade would necessarily have xmin = FrozenTransactionId. +So we don't need tuple header t_xmin field and we reuse t_xmin to store higher +32 bits of its XMAX. + +Double XMAX format is only for full pages that don't have 16 bytes for +pd_special. So it neither has a�place for a single tuple. Insert and HOT update +for double XMAX pages is impossible and not supported. We can only read or +delete tuples from it. + +When we are able to prune page double XMAX it will be converted from it to +general 64-bit XID page format with all operations on its tuples supported. + +In-memory tuple format +---------------------- + +In-memory tuple representation consists of two parts: +- HeapTupleHeader from disk page (contains all heap tuple contents, not only +header) +- HeapTuple with additional in-memory fields + +HeapTuple for each tuple in memory stores t_xid_base/t_multi_base - a copies of +page's pd_xid_base/pd_multi_base. With tuple's 32-bit t_xmin and t_xmax from +HeapTupleHeader they are used to calculate actual 64-bit XMIN and XMAX: + +XMIN = t_xmin + t_xid_base. (3) +XMAX = t_xmax + t_xid_base/t_multi_base. (4) + +The downside of this is that we can not use tuple's XMIN and XMAX right away. +We often need to re-read t_xmin and t_xmax - which could actually be pointers +into a page in shared buffers and therefore they could be updated by any other +backend. + +Update/delete with 64-bit XIDs and 32-bit t_xmin/t_xmax +-------------------------------------------------------------- + +When we try to delete/update a tuple, we check that XMAX for a page fits (2). +I.e. that t_xmax will not be over MaxShortTransactionId relative to +pd_xid_base/pd_multi_base of a its page. + +If the current XID doesn't fit a range +(pd_xid_base, pd_xid_base + MaxShortTransactionId) (5): + +- heap_page_prepare_for_xid() will try to increase pd_xid_base/pd_multi_base on +a page and update all t_xmin/t_xmax of the other tuples on the page to +correspond new pd_xid_base/pd_multi_base. + +- If it was impossible, it will try to prune and freeze tuples on a page. + +- If this is unsuccessful it will throw an error. Normally this is very +unlikely but if there is a very old living transaction with an age of around +2^32 this can arise. Basically, this is a behavior similar to one during the +vacuum to prevent wraparound when XID was 32-bit. Dba should take care and +avoid very-long-living transactions with an age close to 2^32. So long-living +transactions often they are most likely defunct. + +Insert with 64-bit XIDs and 32-bit t_xmin/t_xmax +------------------------------------------------ + +On insert we check if current XID fits a range (5). Otherwise: + +- heap_page_prepare_for_xid() will try to increase pd_xid_base for t_xmin will +not be over MaxShortTransactionId. + +- If it is impossible, then it will try to prune and freeze tuples on a page. + +Known issue: if pd_xid_base could not be shifted to accommodate a tuple being +inserted due to a very long-running transaction, we just throw an error. We +neither try to insert a�tuple into another page nor mark the current page as +full. So, in this (unlikely) case we will get regular insert errors on the next +tries to insert to the page 'locked' by this very long-running transaction. + +Upgrade from 32-bit XID versions +-------------------------------- + +pg_upgrade doesn't change pages format itself. It is done lazily after. + +1. At first heap page read, tuples on a page are repacked to free 16 bytes +at the end of a page, possibly freeing space from dead tuples. + +2A. 16 bytes of pd_special is added if there is a place for it + +2B. Page is converted to "Double XMAX" format if there is no place for +pd_special + +3. If a page is in double XMAX format after its first read, and vacuum (or +micro-vacuum at select query) could prune some tuples and free space for +pd_special, prune_page will add pd_special and convert page from double XMAX +to general 64-bit XID page format. -- 2.24.3 (Apple Git-128) --cpok4wp6gsarlzvp-- ^ permalink raw reply [nested|flat] 265+ messages in thread
* [PATCH v1 1/2] Support changing a column into a stored generated column @ 2026-03-16 23:25 Alberto Piai <[email protected]> 0 siblings, 0 replies; 265+ messages in thread From: Alberto Piai @ 2026-03-16 23:25 UTC (permalink / raw) This adds basic support for an ALTER TABLE ... ALTER COLUMN command to turn a regular column into a stored generated column. The syntax is chosen to be similar to ... ALTER COLUMN ... ADD GENERATED ... AS IDENTITY with the difference that in this case, since we're dealing with a generated column, only ALWAYS is supported. Additionally, STORED must always be specified. Since this is a first prototype, no thought has been given to partitioned nor foreign tables, so these are not supported either. This operation always rewrites the contents of the column using the new generated expression. --- src/backend/commands/tablecmds.c | 137 +++++++++++++++++- src/backend/parser/gram.y | 31 ++++ src/include/nodes/parsenodes.h | 1 + .../test_ddl_deparse/test_ddl_deparse.c | 2 + src/test/regress/expected/alter_table.out | 122 ++++++++++++++++ src/test/regress/sql/alter_table.sql | 69 +++++++++ 6 files changed, 361 insertions(+), 1 deletion(-) diff --git a/src/backend/commands/tablecmds.c b/src/backend/commands/tablecmds.c index 67e42e5df29..e7386e81b07 100644 --- a/src/backend/commands/tablecmds.c +++ b/src/backend/commands/tablecmds.c @@ -760,6 +760,10 @@ static void ATExecMergePartitions(List **wqueue, AlteredTableInfo *tab, Relation static void ATExecSplitPartition(List **wqueue, AlteredTableInfo *tab, Relation rel, PartitionCmd *cmd, AlterTableUtilityContext *context); +static ObjectAddress ATExecAddGeneratedAsExprStored(AlteredTableInfo *tab, + Relation rel, + const char *colName, + Constraint *def); /* ---------------------------------------------------------------- * DefineRelation @@ -4746,6 +4750,7 @@ AlterTableGetLockLevel(List *cmds) case AT_SetExpression: case AT_DropExpression: case AT_SetCompression: + case AT_AddGeneratedAsExprStored: cmd_lockmode = AccessExclusiveLock; break; @@ -5321,6 +5326,16 @@ ATPrepCmd(List **wqueue, Relation rel, AlterTableCmd *cmd, /* No command-specific prep needed */ pass = AT_PASS_MISC; break; + case AT_AddGeneratedAsExprStored: + /* No support yet for: partitioned tables, foreign tables */ + ATSimplePermissions(cmd->subtype, rel, ATT_TABLE); + + /* + * This has similar mechanics to AT_SetExpression, let's use the + * same pass. + */ + pass = AT_PASS_SET_EXPRESSION; + break; default: /* oops */ elog(ERROR, "unrecognized alter table type: %d", (int) cmd->subtype); @@ -5733,6 +5748,12 @@ ATExecCmd(List **wqueue, AlteredTableInfo *tab, ATExecSplitPartition(wqueue, tab, rel, (PartitionCmd *) cmd->def, context); break; + case AT_AddGeneratedAsExprStored: + cmd = ATParseTransformCmd(wqueue, tab, rel, cmd, false, lockmode, + cur_pass, context); + Assert(cmd != NULL); + address = ATExecAddGeneratedAsExprStored(tab, rel, cmd->name, (Constraint *) cmd->def); + break; default: /* oops */ elog(ERROR, "unrecognized alter table type: %d", (int) cmd->subtype); @@ -6785,6 +6806,8 @@ alter_table_type_to_string(AlterTableType cmdtype) return "ALTER COLUMN ... DROP IDENTITY"; case AT_ReAddStatistics: return NULL; /* not real grammar */ + case AT_AddGeneratedAsExprStored: + return "ALTER COLUMN ... ADD GENERATED ALWAYS AS (...) STORED"; } return NULL; @@ -8823,6 +8846,118 @@ ATExecSetExpression(AlteredTableInfo *tab, Relation rel, const char *colName, return address; } +/* + * ALTER TABLE ALTER COLUMN ADD GENERATED ALWAYS AS expr STORED + */ +static ObjectAddress +ATExecAddGeneratedAsExprStored(AlteredTableInfo *tab, + Relation rel, + const char *colName, + Constraint *def) +{ + HeapTuple tuple; + Form_pg_attribute attTup; + AttrNumber attnum; + ObjectAddress address; + Expr *defval; + NewColumnValue *newval; + RawColumnDefault *rawEnt; + Relation pg_attribute; + + Assert(def->raw_expr != NULL); + Assert(def->cooked_expr == NULL); + Assert(def->generated_when == ATTRIBUTE_IDENTITY_ALWAYS); + Assert(def->generated_kind == ATTRIBUTE_GENERATED_STORED); + + tuple = SearchSysCacheAttName(RelationGetRelid(rel), colName); + if (!HeapTupleIsValid(tuple)) + ereport(ERROR, + (errcode(ERRCODE_UNDEFINED_COLUMN), + errmsg("column \"%s\" of relation \"%s\" does not exist", + colName, RelationGetRelationName(rel)))); + + attTup = (Form_pg_attribute) GETSTRUCT(tuple); + + attnum = attTup->attnum; + if (attnum <= 0) + ereport(ERROR, + (errcode(ERRCODE_FEATURE_NOT_SUPPORTED), + errmsg("cannot alter system column \"%s\"", + colName))); + + if (attTup->attgenerated) + ereport(ERROR, + (errcode(ERRCODE_OBJECT_NOT_IN_PREREQUISITE_STATE), + errmsg("column \"%s\" of relation \"%s\" is already a generated column", + colName, RelationGetRelationName(rel)))); + + /* Mark as generated stored in pg_attribute */ + pg_attribute = table_open(AttributeRelationId, RowExclusiveLock); + attTup->attgenerated = ATTRIBUTE_GENERATED_STORED; + CatalogTupleUpdate(pg_attribute, &tuple->t_self, tuple); + table_close(pg_attribute, RowExclusiveLock); + + /* Make above changes visible */ + CommandCounterIncrement(); + + ReleaseSysCache(tuple); + + /* + * Find everything that depends on the column (constraints, indexes, etc), + * and record enough information to let us recreate the objects. + */ + RememberAllDependentForRebuilding(tab, AT_AddGeneratedAsExprStored, + rel, attnum, colName); + + /* + * Remove previous default value, if any, and store the new generator + * expression. + */ + RemoveAttrDefault(RelationGetRelid(rel), attnum, DROP_RESTRICT, + false, false); + + rawEnt = palloc_object(RawColumnDefault); + rawEnt->attnum = attnum; + rawEnt->raw_default = def->raw_expr; + rawEnt->generated = def->generated_kind; + AddRelationNewConstraints(rel, list_make1(rawEnt), NIL, + false, true, false, NULL); + + /* Make above changes visible */ + CommandCounterIncrement(); + + /* + * Clear all the missing values if we're rewriting the table, since this + * renders them pointless. + */ + RelationClearMissing(rel); + + /* Make above changes visible */ + CommandCounterIncrement(); + + /* Drop any pg_statistic entry for the column */ + RemoveStatistics(RelationGetRelid(rel), attnum); + + /* Build a concrete expression for the new default (generated) value */ + defval = (Expr *) build_column_default(rel, attnum); + defval = expression_planner(defval); + + /* Schedule a rewrite */ + newval = palloc0_object(NewColumnValue); + newval->attnum = attnum; + newval->expr = defval; + newval->is_generated = true; + tab->newvals = lappend(tab->newvals, newval); + tab->rewrite |= AT_REWRITE_DEFAULT_VAL; + + InvokeObjectPostAlterHook(RelationRelationId, + RelationGetRelid(rel), attnum); + + ObjectAddressSubSet(address, RelationRelationId, + RelationGetRelid(rel), attnum); + return address; +} + /* * ALTER TABLE ALTER COLUMN DROP EXPRESSION */ @@ -15290,7 +15425,7 @@ RememberAllDependentForRebuilding(AlteredTableInfo *tab, AlterTableType subtype, SysScanDesc scan; HeapTuple depTup; - Assert(subtype == AT_AlterColumnType || subtype == AT_SetExpression); + Assert(subtype == AT_AlterColumnType || subtype == AT_SetExpression || subtype == AT_AddGeneratedAsExprStored); depRel = table_open(DependRelationId, RowExclusiveLock); diff --git a/src/backend/parser/gram.y b/src/backend/parser/gram.y index c2584249603..74440e801d4 100644 --- a/src/backend/parser/gram.y +++ b/src/backend/parser/gram.y @@ -2717,6 +2717,37 @@ alter_table_cmd: n->name = $3; n->def = (Node *) c; + $$ = (Node *) n; + } + /* ALTER TABLE <name> ALTER [COLUMN] <colname> ADD GENERATED ALWAYS AS ( <expression> ) STORED */ + | ALTER opt_column ColId ADD_P GENERATED generated_when AS '(' a_expr ')' STORED + { + AlterTableCmd *n = makeNode(AlterTableCmd); + Constraint *c = makeNode(Constraint); + + c->contype = CONSTR_GENERATED; + c->generated_when = $6; + c->raw_expr = $9; + c->cooked_expr = NULL; + c->generated_kind = ATTRIBUTE_GENERATED_STORED; + c->location = @5; + + /* + * Like in the case of ColConstraintElem, we cannot handle + * this in the grammar because IDENTITY allows both ALWAYS + * and BY DEFAULT, while generated columns only allow + * ALWAYS. This would lead to shift/reduce conflicts. + */ + if (c->generated_when != ATTRIBUTE_IDENTITY_ALWAYS) + ereport(ERROR, + (errcode(ERRCODE_SYNTAX_ERROR), + errmsg("for a generated column, GENERATED ALWAYS must be specified"), + parser_errposition(@6))); + + n->subtype = AT_AddGeneratedAsExprStored; + n->name = $3; + n->def = (Node *) c; + $$ = (Node *) n; } /* ALTER TABLE <name> ALTER [COLUMN] <colname> SET <sequence options>/RESET */ diff --git a/src/include/nodes/parsenodes.h b/src/include/nodes/parsenodes.h index ffadd667167..6b61513e6d0 100644 --- a/src/include/nodes/parsenodes.h +++ b/src/include/nodes/parsenodes.h @@ -2568,6 +2568,7 @@ typedef enum AlterTableType AT_SetIdentity, /* SET identity column options */ AT_DropIdentity, /* DROP IDENTITY */ AT_ReAddStatistics, /* internal to commands/tablecmds.c */ + AT_AddGeneratedAsExprStored, /* ADD GENERATED ALWAYS AS (...) STORED */ } AlterTableType; typedef struct AlterTableCmd /* one subcommand of an ALTER TABLE */ diff --git a/src/test/modules/test_ddl_deparse/test_ddl_deparse.c b/src/test/modules/test_ddl_deparse/test_ddl_deparse.c index 64a1dfa9f79..7c1699d538c 100644 --- a/src/test/modules/test_ddl_deparse/test_ddl_deparse.c +++ b/src/test/modules/test_ddl_deparse/test_ddl_deparse.c @@ -315,6 +315,8 @@ get_altertable_subcmdinfo(PG_FUNCTION_ARGS) case AT_ReAddStatistics: strtype = "(re) ADD STATS"; break; + case AT_AddGeneratedAsExprStored: + strtype = "ADD GENERATED ALWAYS AS (...) STORED"; } if (subcmd->recurse) diff --git a/src/test/regress/expected/alter_table.out b/src/test/regress/expected/alter_table.out index ccd79dfecc0..75f64628aef 100644 --- a/src/test/regress/expected/alter_table.out +++ b/src/test/regress/expected/alter_table.out @@ -4863,3 +4863,125 @@ drop publication pub1; drop schema alter1 cascade; drop schema alter2 cascade; NOTICE: drop cascades to table alter2.t1 +-- Tests for ALTER COLUMN ... ADD GENERATED ALWAYS as ( expr ) STORED +-- turning a regular column into a stored generated column +create schema testgen; +create table testgen.t1 (a int, b int not null); +insert into testgen.t1 (a, b) + select x, x from generate_series(1, 10) x; +alter table testgen.t1 alter column b + add generated always as (a * 2) stored; +\d+ testgen.t1 + Table "testgen.t1" + Column | Type | Collation | Nullable | Default | Storage | Stats target | Description +--------+---------+-----------+----------+------------------------------------+---------+--------------+------------- + a | integer | | | | plain | | + b | integer | | not null | generated always as (a * 2) stored | plain | | +Not-null constraints: + "t1_b_not_null" NOT NULL "b" + +select a, b, a * 2 as expected, b = (a * 2) as correct + from testgen.t1 order by a; + a | b | expected | correct +----+----+----------+--------- + 1 | 2 | 2 | t + 2 | 4 | 4 | t + 3 | 6 | 6 | t + 4 | 8 | 8 | t + 5 | 10 | 10 | t + 6 | 12 | 12 | t + 7 | 14 | 14 | t + 8 | 16 | 16 | t + 9 | 18 | 18 | t + 10 | 20 | 20 | t +(10 rows) + +insert into testgen.t1 (a, b) values (10, 20); +ERROR: cannot insert a non-DEFAULT value into column "b" +DETAIL: Column "b" is a generated column. +insert into testgen.t1 (a, b) values (10, 21); +ERROR: cannot insert a non-DEFAULT value into column "b" +DETAIL: Column "b" is a generated column. +drop table testgen.t1; +-- turning a regular column into a stored generated column +-- fails when another constraint conflicts with the new expression +create table testgen.t2 (a int, b int not null); +insert into testgen.t2 (a, b) select x, x * 2 from generate_series(0, 5) x; +alter table testgen.t2 add constraint chk_gen_clause check (b = a * 2); +select pg_relation_filenode('testgen.t2') as t2_filenode_before \gset +alter table testgen.t2 alter column b add generated always as (a * 3) stored; +ERROR: check constraint "chk_gen_clause" of relation "t2" is violated by some row +select pg_relation_filenode('testgen.t2') as t2_filenode_after \gset +select :t2_filenode_before = :t2_filenode_after as did_not_rewrite; + did_not_rewrite +----------------- + t +(1 row) + +\d+ testgen.t2 + Table "testgen.t2" + Column | Type | Collation | Nullable | Default | Storage | Stats target | Description +--------+---------+-----------+----------+---------+---------+--------------+------------- + a | integer | | | | plain | | + b | integer | | not null | | plain | | +Check constraints: + "chk_gen_clause" CHECK (b = (a * 2)) +Not-null constraints: + "t2_b_not_null" NOT NULL "b" + +drop table testgen.t2; +-- rewrite an indexed column +create table testgen.t3 (a int, b int); +create index idx_b on testgen.t3 (b); +insert into testgen.t3 (a, b) select x, x from generate_series(1, 10) x; +select pg_relation_filenode('testgen.idx_b') as idx_filenode_before \gset +alter table testgen.t3 alter column b add generated always as (a * 2) stored; +select pg_relation_filenode('testgen.idx_b') as idx_filenode_after \gset +select :idx_filenode_before != :idx_filenode_after as did_rewrite_idx; + did_rewrite_idx +----------------- + t +(1 row) + +-- tests for invalid invocations +alter table doesnotexist alter column foo + add generated always as (bar * 2) stored; +ERROR: relation "doesnotexist" does not exist +create table testgen.t1 (a int); +alter table testgen.t1 alter column doesnotexist + add generated always as (bar * 2) stored; +ERROR: column "doesnotexist" of relation "t1" does not exist +alter table testgen.t1 add column b int; +alter table testgen.t1 alter column b + add generated always as (doesnotexist * 2) stored; +ERROR: column "doesnotexist" does not exist +-- invalid: only supports ALWAYS +alter table testgen.t1 alter column b + add generated by default as (a * 2) stored; +ERROR: for a generated column, GENERATED ALWAYS must be specified +LINE 2: add generated by default as (a * 2) stored; + ^ +-- invalid: only supports STORED +alter table testgen.t1 alter column b add generated always as (a * 2); +ERROR: syntax error at or near ";" +LINE 1: ...e testgen.t1 alter column b add generated always as (a * 2); + ^ +alter table testgen.t1 alter column b add generated always as (a * 2) virtual; +ERROR: syntax error at or near "virtual" +LINE 1: ...n.t1 alter column b add generated always as (a * 2) virtual; + ^ +-- invalid: b is already a generated column +create table testgen.t2 (a int, b int generated always as (a * 2) stored); +alter table testgen.t2 alter column b add generated always as (a * 2) stored; +ERROR: column "b" of relation "t2" is already a generated column +-- not supported: partitioned tables +create table testgen.tpart (a int, b int) partition by hash (a); +alter table testgen.tpart alter column b add generated always as (a * 2) stored; +ERROR: ALTER action ALTER COLUMN ... ADD GENERATED ALWAYS AS (...) STORED cannot be performed on relation "tpart" +DETAIL: This operation is not supported for partitioned tables. +drop schema testgen cascade; +NOTICE: drop cascades to 4 other objects +DETAIL: drop cascades to table testgen.t3 +drop cascades to table testgen.t1 +drop cascades to table testgen.t2 +drop cascades to table testgen.tpart diff --git a/src/test/regress/sql/alter_table.sql b/src/test/regress/sql/alter_table.sql index f5f13bbd3e7..d776595a6ed 100644 --- a/src/test/regress/sql/alter_table.sql +++ b/src/test/regress/sql/alter_table.sql @@ -3159,3 +3159,72 @@ alter table alter1.t1 set schema alter2; drop publication pub1; drop schema alter1 cascade; drop schema alter2 cascade; + +-- Tests for ALTER COLUMN ... ADD GENERATED ALWAYS as ( expr ) STORED +-- turning a regular column into a stored generated column +create schema testgen; + +create table testgen.t1 (a int, b int not null); +insert into testgen.t1 (a, b) + select x, x from generate_series(1, 10) x; +alter table testgen.t1 alter column b + add generated always as (a * 2) stored; +\d+ testgen.t1 +select a, b, a * 2 as expected, b = (a * 2) as correct + from testgen.t1 order by a; +insert into testgen.t1 (a, b) values (10, 20); +insert into testgen.t1 (a, b) values (10, 21); +drop table testgen.t1; + +-- turning a regular column into a stored generated column +-- fails when another constraint conflicts with the new expression +create table testgen.t2 (a int, b int not null); +insert into testgen.t2 (a, b) select x, x * 2 from generate_series(0, 5) x; +alter table testgen.t2 add constraint chk_gen_clause check (b = a * 2); +select pg_relation_filenode('testgen.t2') as t2_filenode_before \gset +alter table testgen.t2 alter column b add generated always as (a * 3) stored; +select pg_relation_filenode('testgen.t2') as t2_filenode_after \gset +select :t2_filenode_before = :t2_filenode_after as did_not_rewrite; +\d+ testgen.t2 +drop table testgen.t2; + +-- rewrite an indexed column +create table testgen.t3 (a int, b int); +create index idx_b on testgen.t3 (b); +insert into testgen.t3 (a, b) select x, x from generate_series(1, 10) x; +select pg_relation_filenode('testgen.idx_b') as idx_filenode_before \gset +alter table testgen.t3 alter column b add generated always as (a * 2) stored; +select pg_relation_filenode('testgen.idx_b') as idx_filenode_after \gset +select :idx_filenode_before != :idx_filenode_after as did_rewrite_idx; + +-- tests for invalid invocations +alter table doesnotexist alter column foo + add generated always as (bar * 2) stored; + +create table testgen.t1 (a int); + +alter table testgen.t1 alter column doesnotexist + add generated always as (bar * 2) stored; + +alter table testgen.t1 add column b int; + +alter table testgen.t1 alter column b + add generated always as (doesnotexist * 2) stored; + +-- invalid: only supports ALWAYS +alter table testgen.t1 alter column b + add generated by default as (a * 2) stored; + +-- invalid: only supports STORED +alter table testgen.t1 alter column b add generated always as (a * 2); +alter table testgen.t1 alter column b add generated always as (a * 2) virtual; + +-- invalid: b is already a generated column +create table testgen.t2 (a int, b int generated always as (a * 2) stored); +alter table testgen.t2 alter column b add generated always as (a * 2) stored; + +-- not supported: partitioned tables +create table testgen.tpart (a int, b int) partition by hash (a); +alter table testgen.tpart alter column b add generated always as (a * 2) stored; + +drop schema testgen cascade; -- 2.51.2 --24dzbv6kpqxe4yje Content-Type: text/x-patch; charset=utf-8 Content-Disposition: attachment; filename="v1-0002-Try-to-avoid-a-rewrite-when-adding-a-stored-gener.patch" ^ permalink raw reply [nested|flat] 265+ messages in thread
end of thread, other threads:[~2026-03-16 23:25 UTC | newest] Thread overview: 265+ messages (download: mbox mbox.gz follow: Atom feed) -- links below jump to the message on this page -- 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2022-01-10 19:20 [PATCH v6] README for 64bit xid Pavel Borisov <[email protected]> 2026-03-16 23:25 [PATCH v1 1/2] Support changing a column into a stored generated column Alberto Piai <[email protected]>
This inbox is served by agora; see mirroring instructions for how to clone and mirror all data and code used for this inbox