($INBOX_DIR/description missing)
help / color / mirror / Atom feed[PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
21+ messages / 2 participants
[nested] [flat]
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots
@ 2020-06-30 12:09 Kyotaro Horiguchi <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: Kyotaro Horiguchi @ 2020-06-30 12:09 UTC (permalink / raw)
pg_replication_slot.min_safe_lsn, which shows the oldest LSN kept in
pg_wal, is doubtful in usability for monitoring. Change it to
distance, which shows how many bytes the server can advance before the
slot loses required segments.
---
src/backend/access/transam/xlog.c | 39 +++++++++++++++++++++++
src/backend/catalog/system_views.sql | 2 +-
src/backend/replication/slotfuncs.c | 19 +++++------
src/include/access/xlog.h | 1 +
src/include/catalog/pg_proc.dat | 4 +--
src/test/recovery/t/019_replslot_limit.pl | 20 ++++++------
src/test/regress/expected/rules.out | 4 +--
7 files changed, 62 insertions(+), 27 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index fd93bcfaeb..1f27639912 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -9570,6 +9570,45 @@ GetWALAvailability(XLogRecPtr targetLSN)
return WALAVAIL_REMOVED;
}
+/*
+ * Calculate how many bytes we can advance from currptr until the targetLSN is
+ * removed.
+ *
+ * Returns 0 if the distance is invalid.
+ */
+uint64
+DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr)
+{
+ XLogSegNo targetSeg;
+ XLogSegNo keepSegs;
+ XLogSegNo failSeg;
+ XLogRecPtr horizon;
+
+ XLByteToSeg(targetLSN, targetSeg, wal_segment_size);
+ keepSegs = 0;
+
+ /* no limit if max_slot_wal_keep_size is invalid */
+ if (max_slot_wal_keep_size_mb < 0)
+ return 0;
+
+ /* How many segments slots can keep? */
+ keepSegs = ConvertToXSegs(max_slot_wal_keep_size_mb, wal_segment_size);
+
+ /* override by wal_keep_segments if needed */
+ if (wal_keep_segments > keepSegs)
+ keepSegs = wal_keep_segments;
+
+ /* calculate the LSN where targetLSN is lost when currpos reaches */
+ failSeg = targetSeg + keepSegs + 1;
+ XLogSegNoOffsetToRecPtr(failSeg, 0, wal_segment_size, horizon);
+
+ /* If currptr already beyond the horizon, return zero. */
+ if (currptr > horizon)
+ return 0;
+
+ /* return the distance from currptr to the horizon */
+ return horizon - currptr;
+}
/*
* Retreat *logSegNo to the last segment that we need to retain because of
diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql
index 5314e9348f..b9847a9f92 100644
--- a/src/backend/catalog/system_views.sql
+++ b/src/backend/catalog/system_views.sql
@@ -879,7 +879,7 @@ CREATE VIEW pg_replication_slots AS
L.restart_lsn,
L.confirmed_flush_lsn,
L.wal_status,
- L.min_safe_lsn
+ L.distance
FROM pg_get_replication_slots() AS L
LEFT JOIN pg_database D ON (L.datoid = D.oid);
diff --git a/src/backend/replication/slotfuncs.c b/src/backend/replication/slotfuncs.c
index 88033a79b2..532b3c5826 100644
--- a/src/backend/replication/slotfuncs.c
+++ b/src/backend/replication/slotfuncs.c
@@ -242,6 +242,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Tuplestorestate *tupstore;
MemoryContext per_query_ctx;
MemoryContext oldcontext;
+ XLogRecPtr currlsn;
int slotno;
/* check to see if caller supports us returning a tuplestore */
@@ -274,6 +275,8 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
MemoryContextSwitchTo(oldcontext);
+ currlsn = GetXLogWriteRecPtr();
+
LWLockAcquire(ReplicationSlotControlLock, LW_SHARED);
for (slotno = 0; slotno < max_replication_slots; slotno++)
{
@@ -282,7 +285,7 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
Datum values[PG_GET_REPLICATION_SLOTS_COLS];
bool nulls[PG_GET_REPLICATION_SLOTS_COLS];
WALAvailability walstate;
- XLogSegNo last_removed_seg;
+ uint32 distance;
int i;
if (!slot->in_use)
@@ -398,16 +401,10 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
break;
}
- if (max_slot_wal_keep_size_mb >= 0 &&
- (walstate == WALAVAIL_RESERVED || walstate == WALAVAIL_EXTENDED) &&
- ((last_removed_seg = XLogGetLastRemovedSegno()) != 0))
- {
- XLogRecPtr min_safe_lsn;
-
- XLogSegNoOffsetToRecPtr(last_removed_seg + 1, 0,
- wal_segment_size, min_safe_lsn);
- values[i++] = Int64GetDatum(min_safe_lsn);
- }
+ distance =
+ DistanceToWALHorizon(slot_contents.data.restart_lsn, currlsn);
+ if (distance > 0)
+ values[i++] = Int64GetDatum(distance);
else
nulls[i++] = true;
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index 77ac4e785f..1ec448c5d5 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -329,6 +329,7 @@ extern void InitXLOGAccess(void);
extern void CreateCheckPoint(int flags);
extern bool CreateRestartPoint(int flags);
extern WALAvailability GetWALAvailability(XLogRecPtr targetLSN);
+extern uint64 DistanceToWALHorizon(XLogRecPtr targetLSN, XLogRecPtr currptr);
extern XLogRecPtr CalculateMaxmumSafeLSN(void);
extern void XLogPutNextOid(Oid nextOid);
extern XLogRecPtr XLogRestorePoint(const char *rpName);
diff --git a/src/include/catalog/pg_proc.dat b/src/include/catalog/pg_proc.dat
index 61f2c2f5b4..199fd994bd 100644
--- a/src/include/catalog/pg_proc.dat
+++ b/src/include/catalog/pg_proc.dat
@@ -10063,9 +10063,9 @@
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
proretset => 't', provolatile => 's', prorettype => 'record',
proargtypes => '',
- proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,pg_lsn}',
+ proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8}',
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o}',
- proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,min_safe_lsn}',
+ proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,distance}',
prosrc => 'pg_get_replication_slots' },
{ oid => '3786', descr => 'set up a logical replication slot',
proname => 'pg_create_logical_replication_slot', provolatile => 'v',
diff --git a/src/test/recovery/t/019_replslot_limit.pl b/src/test/recovery/t/019_replslot_limit.pl
index 7d22ae5720..1c76d2d9e9 100644
--- a/src/test/recovery/t/019_replslot_limit.pl
+++ b/src/test/recovery/t/019_replslot_limit.pl
@@ -28,7 +28,7 @@ $node_master->safe_psql('postgres',
# The slot state and remain should be null before the first connection
my $result = $node_master->safe_psql('postgres',
- "SELECT restart_lsn IS NULL, wal_status is NULL, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT restart_lsn IS NULL, wal_status is NULL, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "t|t|t", 'check the state of non-reserved slot is "unknown"');
@@ -52,9 +52,9 @@ $node_master->wait_for_catchup($node_standby, 'replay', $start_lsn);
# Stop standby
$node_standby->stop;
-# Preparation done, the slot is the state "normal" now
+# Preparation done, the slot is the state "reserved" now
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check the catching-up state');
@@ -64,7 +64,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when fitting max_wal_size
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t",
'check that it is safe if WAL fits in max_wal_size');
@@ -74,7 +74,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
# The slot is always "safe" when max_slot_wal_keep_size is not set
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "reserved|t", 'check that slot is working');
@@ -94,9 +94,7 @@ max_slot_wal_keep_size = ${max_slot_wal_keep_size_mb}MB
));
$node_master->reload;
-# The slot is in safe state. The distance from the min_safe_lsn should
-# be as almost (max_slot_wal_keep_size - 1) times large as the segment
-# size
+# The slot is in safe state.
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
@@ -110,7 +108,7 @@ $node_master->safe_psql('postgres', "CHECKPOINT;");
$result = $node_master->safe_psql('postgres',
"SELECT wal_status FROM pg_replication_slots WHERE slot_name = 'rep1'");
is($result, "reserved",
- 'check that min_safe_lsn gets close to the current LSN');
+ 'check that distance gets close to the current LSN');
# The standby can reconnect to master
$node_standby->start;
@@ -154,7 +152,7 @@ advance_wal($node_master, 1);
# Slot gets into 'unreserved' state
$result = $node_master->safe_psql('postgres',
- "SELECT wal_status, min_safe_lsn is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT wal_status, distance is NULL FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "unreserved|t",
'check that the slot state changes to "unreserved"');
@@ -186,7 +184,7 @@ ok( find_in_log(
# This slot should be broken
$result = $node_master->safe_psql('postgres',
- "SELECT slot_name, active, restart_lsn IS NULL, wal_status, min_safe_lsn FROM pg_replication_slots WHERE slot_name = 'rep1'"
+ "SELECT slot_name, active, restart_lsn IS NULL, wal_status, distance FROM pg_replication_slots WHERE slot_name = 'rep1'"
);
is($result, "rep1|f|t|lost|",
'check that the slot became inactive and the state "lost" persists');
diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out
index b813e32215..392eab12dd 100644
--- a/src/test/regress/expected/rules.out
+++ b/src/test/regress/expected/rules.out
@@ -1464,8 +1464,8 @@ pg_replication_slots| SELECT l.slot_name,
l.restart_lsn,
l.confirmed_flush_lsn,
l.wal_status,
- l.min_safe_lsn
- FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, min_safe_lsn)
+ l.distance
+ FROM (pg_get_replication_slots() l(slot_name, plugin, slot_type, datoid, temporary, active, active_pid, xmin, catalog_xmin, restart_lsn, confirmed_flush_lsn, wal_status, distance)
LEFT JOIN pg_database d ON ((l.datoid = d.oid)));
pg_roles| SELECT pg_authid.rolname,
pg_authid.rolsuper,
--
2.18.4
----Next_Part(Wed_Jul__1_10_32_59_2020_161)----
^ permalink raw reply [nested|flat] 21+ messages in thread
* confirmed flush lsn seems to be move backward in certain error cases
@ 2024-02-16 12:09 vignesh C <[email protected]>
0 siblings, 0 replies; 21+ messages in thread
From: vignesh C @ 2024-02-16 12:09 UTC (permalink / raw)
To: PostgreSQL Hackers <[email protected]>
Hi,
The following assertion failure was seen while testing one scenario
for other patch:
TRAP: failed Assert("s->data.confirmed_flush >=
s->last_saved_confirmed_flush"), File: "slot.c", Line: 1760, PID:
545314
postgres: checkpointer performing shutdown
checkpoint(ExceptionalCondition+0xbb)[0x564ee6870c58]
postgres: checkpointer performing shutdown
checkpoint(CheckPointReplicationSlots+0x18e)[0x564ee65e9c71]
postgres: checkpointer performing shutdown checkpoint(+0x1e1403)[0x564ee61be403]
postgres: checkpointer performing shutdown
checkpoint(CreateCheckPoint+0x78a)[0x564ee61bdace]
postgres: checkpointer performing shutdown
checkpoint(ShutdownXLOG+0x150)[0x564ee61bc735]
postgres: checkpointer performing shutdown checkpoint(+0x5ae28c)[0x564ee658b28c]
postgres: checkpointer performing shutdown
checkpoint(CheckpointerMain+0x31e)[0x564ee658ad55]
postgres: checkpointer performing shutdown
checkpoint(AuxiliaryProcessMain+0x1d1)[0x564ee65888d9]
postgres: checkpointer performing shutdown checkpoint(+0x5b7200)[0x564ee6594200]
postgres: checkpointer performing shutdown
checkpoint(PostmasterMain+0x14da)[0x564ee658f12f]
postgres: checkpointer performing shutdown checkpoint(+0x464fc6)[0x564ee6441fc6]
/lib/x86_64-linux-gnu/libc.so.6(+0x29d90)[0x7ff6afa29d90]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0x80)[0x7ff6afa29e40]
postgres: checkpointer performing shutdown
checkpoint(_start+0x25)[0x564ee60b8e05]
I was able to reproduce this issue with the following steps:
-- Setup
-- Publisher node:
create table t1(c1 int);
create table t2(c1 int);
create publication pub1 for table t1;
create publication pub2 for table t2;
-- Subscriber node:
create table t1(c1 int);
create table t2(c1 int);
create subscription test1 connection 'dbname=postgres host=localhost
port=5432' publication pub1, pub2;
select * from pg_subscription;
-- Actual test
insert into t1 values(10);
insert into t2 values(20);
select pg_sleep(10);
drop publication pub2;
insert into t1 values(10);
insert into t2 values(20);
Stop the publisher to see the assertion.
For me the issue reproduces about twice in five times using the
assert_failure.sh script attached.
After the insert operation is replicated to the subscriber, the
subscriber will set the lsn value sent by the publisher in the
replication origin (in my case it was 0/1510978). publisher will then
send keepalive messages with the current WAL position in the publisher
(in my case it was 0/15109B0), but subscriber will simply send this
position as the flush_lsn to the publisher as there are no ongoing
transactions. Then since the publisher is started, it will identify
that publication does not exist and stop the walsender/apply worker
process. When the apply worker is restarted, we will get the
remote_lsn(in my case it was 0/1510978) of the origin and set it to
origin_startpos. We will start the apply worker with this
origin_startpos (origin's remote_lsn). This position will be sent as
feedback to the walsender process from the below stack:
run_apply_worker->start_apply->LogicalRepApplyLoop->send_feedback.
It will use the following send_feedback function call of
LogicalRepApplyLoop function as in below code here as nothing is
received from walsender:
LogicalRepApplyLoop function
.......
len = walrcv_receive(LogRepWorkerWalRcvConn, &buf, &fd);
if (len != 0)
{
/* Loop to process all available data (without blocking). */
for (;;)
{
CHECK_FOR_INTERRUPTS();
...
}
}
/* confirm all writes so far */
send_feedback(last_received, false, false);
.......
In send_feedback, we will set flushpos to replication origin's
remote_lsn and send it to the walsender process. Walsender process
will receive this information and set confirmed_flush in:
ProcessStandbyReplyMessage->LogicalConfirmReceivedLocation
Then immediately we are trying to stop the publisher instance,
shutdown checkpoint process will be triggered. In this case:
confirmed_flush = 0/1510978 will be lesser than
last_saved_confirmed_flush = 0/15109B0 which will result in Assertion
failure.
This issue is happening because we allow setting the confirmed_flush
to a backward position.
There are a couple of ways to fix this:
a) One way it not to update the confirm_flush if the lsn sent is an
older value like in Confirm_flush_dont_allow_backward.patch
b) Another way is to remove the assertion in
CheckPointReplicationSlots and marking the slot as dirty only if
confirmed_flush is greater than last_saved_confirmed_flush like in
Assert_confirmed_flush_will_always_not_be_less_than_last_saved_confirmed_flush.patch
I preferred the first approach.
Thoughts?
Regards,
Vignesh
Attachments:
[application/x-shellscript] assert_failure.sh (1.5K, ../../CALDaNm3hgow2+oEov5jBk4iYP5eQrUCF1yZtW7+dV3J__p4KLQ@mail.gmail.com/2-assert_failure.sh)
download
[text/x-patch] Confirm_flush_dont_allow_backward.patch (984B, ../../CALDaNm3hgow2+oEov5jBk4iYP5eQrUCF1yZtW7+dV3J__p4KLQ@mail.gmail.com/3-Confirm_flush_dont_allow_backward.patch)
download | inline diff:
diff --git a/src/backend/replication/logical/logical.c b/src/backend/replication/logical/logical.c
index a53815f2ed..98aa629ced 100644
--- a/src/backend/replication/logical/logical.c
+++ b/src/backend/replication/logical/logical.c
@@ -1839,7 +1839,8 @@ LogicalConfirmReceivedLocation(XLogRecPtr lsn)
SpinLockAcquire(&MyReplicationSlot->mutex);
- MyReplicationSlot->data.confirmed_flush = lsn;
+ if (lsn > MyReplicationSlot->data.confirmed_flush)
+ MyReplicationSlot->data.confirmed_flush = lsn;
/* if we're past the location required for bumping xmin, do so */
if (MyReplicationSlot->candidate_xmin_lsn != InvalidXLogRecPtr &&
@@ -1904,7 +1905,8 @@ LogicalConfirmReceivedLocation(XLogRecPtr lsn)
else
{
SpinLockAcquire(&MyReplicationSlot->mutex);
- MyReplicationSlot->data.confirmed_flush = lsn;
+ if (lsn > MyReplicationSlot->data.confirmed_flush)
+ MyReplicationSlot->data.confirmed_flush = lsn;
SpinLockRelease(&MyReplicationSlot->mutex);
}
}
[text/x-patch] Assert_confirmed_flush_will_always_not_be_less_than_last_saved_confirmed_flush.patch (596B, ../../CALDaNm3hgow2+oEov5jBk4iYP5eQrUCF1yZtW7+dV3J__p4KLQ@mail.gmail.com/4-Assert_confirmed_flush_will_always_not_be_less_than_last_saved_confirmed_flush.patch)
download | inline diff:
diff --git a/src/backend/replication/slot.c b/src/backend/replication/slot.c
index 2180a38063..890dd33892 100644
--- a/src/backend/replication/slot.c
+++ b/src/backend/replication/slot.c
@@ -1757,10 +1757,8 @@ CheckPointReplicationSlots(bool is_shutdown)
{
SpinLockAcquire(&s->mutex);
- Assert(s->data.confirmed_flush >= s->last_saved_confirmed_flush);
-
if (s->data.invalidated == RS_INVAL_NONE &&
- s->data.confirmed_flush != s->last_saved_confirmed_flush)
+ s->data.confirmed_flush > s->last_saved_confirmed_flush)
{
s->just_dirtied = true;
s->dirty = true;
^ permalink raw reply [nested|flat] 21+ messages in thread
end of thread, other threads:[~2024-02-16 12:09 UTC | newest]
Thread overview: 21+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2020-06-30 12:09 [PATCH v1] Replace min_safe_lsn with "distance" in pg_replication_slots Kyotaro Horiguchi <[email protected]>
2024-02-16 12:09 confirmed flush lsn seems to be move backward in certain error cases vignesh C <[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