agora inbox for [email protected]  
help / color / mirror / Atom feed
Re: Allow interrupts on waiting standby
23+ messages / 4 participants
[nested] [flat]

* Re: Allow interrupts on waiting standby
@ 2017-03-07 20:45  Peter Eisentraut <[email protected]>
  0 siblings, 1 reply; 23+ messages in thread

From: Peter Eisentraut @ 2017-03-07 20:45 UTC (permalink / raw)
  To: Michael Paquier <[email protected]>; Simon Riggs <[email protected]>; +Cc: Andres Freund <[email protected]>; Robert Haas <[email protected]>; pgsql-hackers

On 1/30/17 20:34, Michael Paquier wrote:
> On Fri, Jan 27, 2017 at 10:17 PM, Michael Paquier
> <[email protected]> wrote:
>> Two things I forgot in this patch:
>> - documentation for the new wait event
>> - the string for the wait event or this would show up as "???" in
>> pg_stat_activity.
>> There are no default clauses in the pgstat_get_wait_* routines so my
>> compiler is actually complaining...
> 
> Both things are fixed in the new version attached. I have added as
> well this patch to the next commit fest:
> https://commitfest.postgresql.org/13/977/

I'm not clear now what this patch is intending to accomplish, seeing
that the original issue has already been fixed.

-- 
Peter Eisentraut              http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services


-- 
Sent via pgsql-hackers mailing list ([email protected])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers



^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* Re: Allow interrupts on waiting standby
@ 2017-03-13 03:44  Michael Paquier <[email protected]>
  parent: Peter Eisentraut <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Michael Paquier @ 2017-03-13 03:44 UTC (permalink / raw)
  To: Peter Eisentraut <[email protected]>; +Cc: Simon Riggs <[email protected]>; Andres Freund <[email protected]>; Robert Haas <[email protected]>; pgsql-hackers

On Wed, Mar 8, 2017 at 5:45 AM, Peter Eisentraut
<[email protected]> wrote:
> On 1/30/17 20:34, Michael Paquier wrote:
>> Both things are fixed in the new version attached. I have added as
>> well this patch to the next commit fest:
>> https://commitfest.postgresql.org/13/977/
>
> I'm not clear now what this patch is intending to accomplish, seeing
> that the original issue has already been fixed.

This makes ResolveRecoveryConflictWithVirtualXIDs() more responsive if
the process is signaled, as the wait in pg_usleep can go up to 1s if
things remain stuck for a long time.
-- 
Michael


-- 
Sent via pgsql-hackers mailing list ([email protected])
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-hackers



^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* [PATCH v4] fix HOT tuples while scanning for index builds
@ 2020-08-12 18:02  Alvaro Herrera <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Alvaro Herrera @ 2020-08-12 18:02 UTC (permalink / raw)

---
 src/backend/access/heap/heapam_handler.c | 20 ++++++++++++++++++++
 src/backend/access/heap/pruneheap.c      |  5 +++--
 2 files changed, 23 insertions(+), 2 deletions(-)

diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 267a6ee25a..ba44e30035 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -1324,6 +1324,12 @@ heapam_index_build_range_scan(Relation heapRelation,
 		 * buffer continuously while visiting the page, so no pruning
 		 * operation can occur either.
 		 *
+		 * In cases with only ShareUpdateExclusiveLock on the table, it's
+		 * possible for some HOT tuples to appear that we didn't know about
+		 * when we first read the page.  To handle that case, we re-obtain the
+		 * list of root offsets when a HOT tuple points to a root item that we
+		 * don't know about.
+		 *
 		 * Also, although our opinions about tuple liveness could change while
 		 * we scan the page (due to concurrent transaction commits/aborts),
 		 * the chain root locations won't, so this info doesn't need to be
@@ -1625,6 +1631,20 @@ heapam_index_build_range_scan(Relation heapRelation,
 
 			offnum = ItemPointerGetOffsetNumber(&heapTuple->t_self);
 
+			/*
+			 * If a HOT tuple points to a root that we don't know
+			 * about, obtain root items afresh.  If that still fails,
+			 * report it as corruption.
+			 */
+			if (root_offsets[offnum - 1] == InvalidOffsetNumber)
+			{
+				Page	page = BufferGetPage(hscan->rs_cbuf);
+
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_SHARE);
+				heap_get_root_tuples(page, root_offsets);
+				LockBuffer(hscan->rs_cbuf, BUFFER_LOCK_UNLOCK);
+			}
+
 			if (!OffsetNumberIsValid(root_offsets[offnum - 1]))
 				ereport(ERROR,
 						(errcode(ERRCODE_DATA_CORRUPTED),
diff --git a/src/backend/access/heap/pruneheap.c b/src/backend/access/heap/pruneheap.c
index 256df4de10..7e3d44dfd6 100644
--- a/src/backend/access/heap/pruneheap.c
+++ b/src/backend/access/heap/pruneheap.c
@@ -732,7 +732,7 @@ heap_page_prune_execute(Buffer buffer,
  * root_offsets[k - 1] = j.
  *
  * The passed-in root_offsets array must have MaxHeapTuplesPerPage entries.
- * We zero out all unused entries.
+ * Unused entries are filled with InvalidOffsetNumber (zero).
  *
  * The function must be called with at least share lock on the buffer, to
  * prevent concurrent prune operations.
@@ -747,7 +747,8 @@ heap_get_root_tuples(Page page, OffsetNumber *root_offsets)
 	OffsetNumber offnum,
 				maxoff;
 
-	MemSet(root_offsets, 0, MaxHeapTuplesPerPage * sizeof(OffsetNumber));
+	MemSet(root_offsets, InvalidOffsetNumber,
+		   MaxHeapTuplesPerPage * sizeof(OffsetNumber));
 
 	maxoff = PageGetMaxOffsetNumber(page);
 	for (offnum = FirstOffsetNumber; offnum <= maxoff; offnum = OffsetNumberNext(offnum))
-- 
2.20.1


--opJtzjQTFsWo+cga--





^ permalink  raw  reply  [nested|flat] 23+ messages in thread

* RFC: Adding \history [options] [filename] to psql (Snippets and Shared Queries)
@ 2023-06-05 15:50  Kirk Wolak <[email protected]>
  0 siblings, 0 replies; 23+ messages in thread

From: Kirk Wolak @ 2023-06-05 15:50 UTC (permalink / raw)
  To: PostgreSQL Hackers <[email protected]>

Everyone,
  After recently deep diving on some readline features and optimizing my
bash environment to have a static set of "snippets" that I can always
find...

  it takes just a couple of history API calls to add some interesting
features for those  that want them.  The result of adding 3-4 such commands
(all under \history, and with compatible flags):

- Saving your current history without exiting (currently doable as \s
:HISTFILE)
- Reloading your history file (so you can easily share something across
sessions) w/o exiting.
- Stack Loading of specific history (like a shared snippets library, and a
personal snippets library) [clearing your history, then loading them in a
custom order]

  The upside is really about clearly identifying and sharing permanent
snippets, while having that list be editable externally.  Again, bringing
teams online who don't always know the PG way of doing things (Waits,
Locks, Space, High CPU queries, Running Queries).

  My intention is to leverage the way PSQL keeps the Comment above the SQL
with the SQL.
Then I can step backwards searching for "context" markers (Ctrl-R) or
-- <CONTEXT> [F8] {history-search-backward}

  To rip through my snippets

Kirk...
PS: I could do all of this under \s [options] [filename] it's just less
clear...


^ permalink  raw  reply  [nested|flat] 23+ messages in thread


end of thread, other threads:[~2023-06-05 15:50 UTC | newest]

Thread overview: 23+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2017-03-07 20:45 Re: Allow interrupts on waiting standby Peter Eisentraut <[email protected]>
2017-03-13 03:44 ` Michael Paquier <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2020-08-12 18:02 [PATCH v4] fix HOT tuples while scanning for index builds Alvaro Herrera <[email protected]>
2023-06-05 15:50 RFC: Adding \history [options] [filename] to psql (Snippets and Shared Queries) Kirk Wolak <[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