Movatterモバイル変換

[0]ホーム
URL:

Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

 Sign up
Appearance settings

Commite68b6ad

Browse files
author
Amit Kapila
committed
Add slotsync_skip_reason column to pg_replication_slots view.
Introduce a new column, slotsync_skip_reason, in the pg_replication_slotsview. This column records the reason why the last slot synchronization wasskipped. It is primarily relevant for logical replication slots on standbyservers where the 'synced' field is true. The value is NULL whensynchronization succeeds.Author: Shlok Kyal <shlok.kyal.oss@gmail.com>Reviewed-by: shveta malik <shveta.malik@gmail.com>Reviewed-by: Hayato Kuroda <kuroda.hayato@fujitsu.com>Reviewed-by: Ashutosh Sharma <ashu.coek88@gmail.com>Reviewed-by: Hou Zhijie <houzj.fnst@fujitsu.com>Reviewed-by: Amit Kapila <amit.kapila16@gmail.com>Discussion:https://postgr.es/m/CAE9k0PkhfKrTEAsGz4DjOhEj1nQ+hbQVfvWUxNacD38ibW3a1g@mail.gmail.com
1 parent9ccc049 commite68b6ad

File tree

13 files changed

+149
-17
lines changed

13 files changed

+149
-17
lines changed

‎doc/src/sgml/monitoring.sgml‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1665,7 +1665,7 @@ description | Waiting for a newly initialized WAL file to reach durable storage
16651665
</para>
16661666
<para>
16671667
Number of times the slot synchronization is skipped. Slot
1668-
synchronizationoccur only on standby servers and thus this column has
1668+
synchronizationoccurs only on standby servers and thus this column has
16691669
no meaning on the primary server.
16701670
</para>
16711671
</entry>
@@ -1677,7 +1677,7 @@ description | Waiting for a newly initialized WAL file to reach durable storage
16771677
</para>
16781678
<para>
16791679
Time at which last slot synchronization was skipped. Slot
1680-
synchronizationoccur only on standby servers and thus this column has
1680+
synchronizationoccurs only on standby servers and thus this column has
16811681
no meaning on the primary server.
16821682
</para>
16831683
</entry>

‎doc/src/sgml/system-views.sgml‎

Lines changed: 43 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -3102,6 +3102,49 @@ SELECT * FROM pg_locks pl LEFT JOIN pg_prepared_xacts ppx
31023102
</para></entry>
31033103
</row>
31043104

3105+
<row>
3106+
<entry role="catalog_table_entry"><para role="column_definition">
3107+
<structfield>slotsync_skip_reason</structfield><type>text</type>
3108+
</para>
3109+
<para>
3110+
The reason for the last slot synchronization skip. Slot
3111+
synchronization occurs only on standby servers and thus this column has
3112+
no meaning on the primary server. It is relevant mainly for logical slots
3113+
on standby servers whose <structfield>synced</structfield> field is
3114+
<literal>true</literal>. It is <literal>NULL</literal> if slot
3115+
synchronization is successful.
3116+
Possible values are:
3117+
<itemizedlist spacing="compact">
3118+
<listitem>
3119+
<para>
3120+
<literal>wal_or_rows_removed</literal> means that the required WALs or
3121+
catalog rows have already been removed or are at the risk of removal
3122+
from the standby.
3123+
</para>
3124+
</listitem>
3125+
<listitem>
3126+
<para>
3127+
<literal>wal_not_flushed</literal> means that the standby had not
3128+
flushed the WAL corresponding to the position reserved on the failover
3129+
slot.
3130+
</para>
3131+
</listitem>
3132+
<listitem>
3133+
<para>
3134+
<literal>no_consistent_snapshot</literal> means that the standby could
3135+
not build a consistent snapshot to decode WALs from
3136+
<structfield>restart_lsn</structfield>.
3137+
</para>
3138+
</listitem>
3139+
<listitem>
3140+
<para>
3141+
<literal>slot_invalidated</literal> means that the synced slot is
3142+
invalidated.
3143+
</para>
3144+
</listitem>
3145+
</itemizedlist>
3146+
</para></entry>
3147+
</row>
31053148
</tbody>
31063149
</tgroup>
31073150
</table>

‎src/backend/catalog/system_views.sql‎

Lines changed: 2 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1060,7 +1060,8 @@ CREATE VIEW pg_replication_slots AS
10601060
L.conflicting,
10611061
L.invalidation_reason,
10621062
L.failover,
1063-
L.synced
1063+
L.synced,
1064+
L.slotsync_skip_reason
10641065
FROM pg_get_replication_slots()AS L
10651066
LEFT JOIN pg_database DON (L.datoid=D.oid);
10661067

‎src/backend/replication/logical/slotsync.c‎

Lines changed: 38 additions & 5 deletions
Original file line numberDiff line numberDiff line change
@@ -148,6 +148,35 @@ typedef struct RemoteSlot
148148
staticvoidslotsync_failure_callback(intcode,Datumarg);
149149
staticvoidupdate_synced_slots_inactive_since(void);
150150

151+
/*
152+
* Update slot sync skip stats. This function requires the caller to acquire
153+
* the slot.
154+
*/
155+
staticvoid
156+
update_slotsync_skip_stats(SlotSyncSkipReasonskip_reason)
157+
{
158+
ReplicationSlot*slot;
159+
160+
Assert(MyReplicationSlot);
161+
162+
slot=MyReplicationSlot;
163+
164+
/*
165+
* Update the slot sync related stats in pg_stat_replication_slot when a
166+
* slot sync is skipped
167+
*/
168+
if (skip_reason!=SS_SKIP_NONE)
169+
pgstat_report_replslotsync(slot);
170+
171+
/* Update the slot sync skip reason */
172+
if (slot->slotsync_skip_reason!=skip_reason)
173+
{
174+
SpinLockAcquire(&slot->mutex);
175+
slot->slotsync_skip_reason=skip_reason;
176+
SpinLockRelease(&slot->mutex);
177+
}
178+
}
179+
151180
/*
152181
* If necessary, update the local synced slot's metadata based on the data
153182
* from the remote slot.
@@ -170,6 +199,7 @@ update_local_synced_slot(RemoteSlot *remote_slot, Oid remote_dbid,
170199
ReplicationSlot*slot=MyReplicationSlot;
171200
boolupdated_xmin_or_lsn= false;
172201
boolupdated_config= false;
202+
SlotSyncSkipReasonskip_reason=SS_SKIP_NONE;
173203

174204
Assert(slot->data.invalidated==RS_INVAL_NONE);
175205

@@ -188,7 +218,7 @@ update_local_synced_slot(RemoteSlot *remote_slot, Oid remote_dbid,
188218
slot->data.catalog_xmin))
189219
{
190220
/* Update slot sync skip stats */
191-
pgstat_report_replslotsync(slot);
221+
update_slotsync_skip_stats(SS_SKIP_WAL_OR_ROWS_REMOVED);
192222

193223
/*
194224
* This can happen in following situations:
@@ -286,12 +316,15 @@ update_local_synced_slot(RemoteSlot *remote_slot, Oid remote_dbid,
286316
* persisted. See update_and_persist_local_synced_slot().
287317
*/
288318
if (found_consistent_snapshot&& !(*found_consistent_snapshot))
289-
pgstat_report_replslotsync(slot);
319+
skip_reason=SS_SKIP_NO_CONSISTENT_SNAPSHOT;
290320
}
291321

292322
updated_xmin_or_lsn= true;
293323
}
294324

325+
/* Update slot sync skip stats */
326+
update_slotsync_skip_stats(skip_reason);
327+
295328
if (remote_dbid!=slot->data.database||
296329
remote_slot->two_phase!=slot->data.two_phase||
297330
remote_slot->failover!=slot->data.failover||
@@ -696,7 +729,7 @@ synchronize_one_slot(RemoteSlot *remote_slot, Oid remote_dbid)
696729
/* Skip the sync of an invalidated slot */
697730
if (slot->data.invalidated!=RS_INVAL_NONE)
698731
{
699-
pgstat_report_replslotsync(slot);
732+
update_slotsync_skip_stats(SS_SKIP_INVALID);
700733

701734
ReplicationSlotRelease();
702735
returnslot_updated;
@@ -711,7 +744,7 @@ synchronize_one_slot(RemoteSlot *remote_slot, Oid remote_dbid)
711744
*/
712745
if (remote_slot->confirmed_lsn>latestFlushPtr)
713746
{
714-
pgstat_report_replslotsync(slot);
747+
update_slotsync_skip_stats(SS_SKIP_WAL_NOT_FLUSHED);
715748

716749
/*
717750
* Can get here only if GUC 'synchronized_standby_slots' on the
@@ -812,7 +845,7 @@ synchronize_one_slot(RemoteSlot *remote_slot, Oid remote_dbid)
812845
*/
813846
if (remote_slot->confirmed_lsn>latestFlushPtr)
814847
{
815-
pgstat_report_replslotsync(slot);
848+
update_slotsync_skip_stats(SS_SKIP_WAL_NOT_FLUSHED);
816849

817850
/*
818851
* Can get here only if GUC 'synchronized_standby_slots' on the

‎src/backend/replication/slot.c‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -491,6 +491,7 @@ ReplicationSlotCreate(const char *name, bool db_specific,
491491
slot->last_saved_confirmed_flush=InvalidXLogRecPtr;
492492
slot->last_saved_restart_lsn=InvalidXLogRecPtr;
493493
slot->inactive_since=0;
494+
slot->slotsync_skip_reason=SS_SKIP_NONE;
494495

495496
/*
496497
* Create the slot on disk. We haven't actually marked the slot allocated

‎src/backend/replication/slotfuncs.c‎

Lines changed: 17 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -24,6 +24,17 @@
2424
#include"utils/guc.h"
2525
#include"utils/pg_lsn.h"
2626

27+
/*
28+
* Map SlotSyncSkipReason enum values to human-readable names.
29+
*/
30+
staticconstchar*SlotSyncSkipReasonNames[]= {
31+
[SS_SKIP_NONE]="none",
32+
[SS_SKIP_WAL_NOT_FLUSHED]="wal_not_flushed",
33+
[SS_SKIP_WAL_OR_ROWS_REMOVED]="wal_or_rows_removed",
34+
[SS_SKIP_NO_CONSISTENT_SNAPSHOT]="no_consistent_snapshot",
35+
[SS_SKIP_INVALID]="slot_invalidated"
36+
};
37+
2738
/*
2839
* Helper function for creating a new physical replication slot with
2940
* given arguments. Note that this function doesn't release the created
@@ -235,7 +246,7 @@ pg_drop_replication_slot(PG_FUNCTION_ARGS)
235246
Datum
236247
pg_get_replication_slots(PG_FUNCTION_ARGS)
237248
{
238-
#definePG_GET_REPLICATION_SLOTS_COLS20
249+
#definePG_GET_REPLICATION_SLOTS_COLS21
239250
ReturnSetInfo*rsinfo= (ReturnSetInfo*)fcinfo->resultinfo;
240251
XLogRecPtrcurrlsn;
241252
intslotno;
@@ -443,6 +454,11 @@ pg_get_replication_slots(PG_FUNCTION_ARGS)
443454

444455
values[i++]=BoolGetDatum(slot_contents.data.synced);
445456

457+
if (slot_contents.slotsync_skip_reason==SS_SKIP_NONE)
458+
nulls[i++]= true;
459+
else
460+
values[i++]=CStringGetTextDatum(SlotSyncSkipReasonNames[slot_contents.slotsync_skip_reason]);
461+
446462
Assert(i==PG_GET_REPLICATION_SLOTS_COLS);
447463

448464
tuplestore_putvalues(rsinfo->setResult,rsinfo->setDesc,

‎src/backend/utils/activity/pgstat_replslot.c‎

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -115,8 +115,8 @@ pgstat_report_replslotsync(ReplicationSlot *slot)
115115
PgStatShared_ReplSlot*shstatent;
116116
PgStat_StatReplSlotEntry*statent;
117117

118-
/* Slot sync stats are valid only for logical slots on standby. */
119-
Assert(SlotIsLogical(slot));
118+
/* Slot sync stats are valid only forsyncedlogical slots on standby. */
119+
Assert(slot->data.synced);
120120
Assert(RecoveryInProgress());
121121

122122
entry_ref=pgstat_get_entry_ref_locked(PGSTAT_KIND_REPLSLOT,InvalidOid,

‎src/include/catalog/catversion.h‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -57,6 +57,6 @@
5757
*/
5858

5959
/*yyyymmddN */
60-
#defineCATALOG_VERSION_NO202511251
60+
#defineCATALOG_VERSION_NO202511281
6161

6262
#endif

‎src/include/catalog/pg_proc.dat‎

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -11519,9 +11519,9 @@
1151911519
proname => 'pg_get_replication_slots', prorows => '10', proisstrict => 'f',
1152011520
proretset => 't', provolatile => 's', prorettype => 'record',
1152111521
proargtypes => '',
11522-
proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8,bool,pg_lsn,timestamptz,bool,text,bool,bool}',
11523-
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o}',
11524-
proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,safe_wal_size,two_phase,two_phase_at,inactive_since,conflicting,invalidation_reason,failover,synced}',
11522+
proallargtypes => '{name,name,text,oid,bool,bool,int4,xid,xid,pg_lsn,pg_lsn,text,int8,bool,pg_lsn,timestamptz,bool,text,bool,bool,text}',
11523+
proargmodes => '{o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o,o}',
11524+
proargnames => '{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,wal_status,safe_wal_size,two_phase,two_phase_at,inactive_since,conflicting,invalidation_reason,failover,synced,slotsync_skip_reason}',
1152511525
prosrc => 'pg_get_replication_slots' },
1152611526
{ oid => '3786', descr => 'set up a logical replication slot',
1152711527
proname => 'pg_create_logical_replication_slot', provolatile => 'v',

‎src/include/replication/slot.h‎

Lines changed: 30 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -71,6 +71,24 @@ typedef enum ReplicationSlotInvalidationCause
7171
/* Maximum number of invalidation causes */
7272
#defineRS_INVAL_MAX_CAUSES 4
7373

74+
/*
75+
* When the slot synchronization worker is running, or when
76+
* pg_sync_replication_slots is executed, slot synchronization may be
77+
* skipped. This enum defines the possible reasons for skipping slot
78+
* synchronization.
79+
*/
80+
typedefenumSlotSyncSkipReason
81+
{
82+
SS_SKIP_NONE,/* No skip */
83+
SS_SKIP_WAL_NOT_FLUSHED,/* Standby did not flush the wal corresponding
84+
* to confirmed flush of remote slot */
85+
SS_SKIP_WAL_OR_ROWS_REMOVED,/* Remote slot is behind; required WAL or
86+
* rows may be removed or at risk */
87+
SS_SKIP_NO_CONSISTENT_SNAPSHOT,/* Standby could not build a consistent
88+
* snapshot */
89+
SS_SKIP_INVALID/* Local slot is invalid */
90+
}SlotSyncSkipReason;
91+
7492
/*
7593
* On-Disk data of a replication slot, preserved across restarts.
7694
*/
@@ -249,6 +267,18 @@ typedef struct ReplicationSlot
249267
*/
250268
XLogRecPtrlast_saved_restart_lsn;
251269

270+
/*
271+
* Reason for the most recent slot synchronization skip.
272+
*
273+
* Slot sync skips can occur for both temporary and persistent replication
274+
* slots. They are more common for temporary slots, but persistent slots
275+
* may also skip synchronization in rare cases (e.g.,
276+
* SS_SKIP_WAL_NOT_FLUSHED or SS_SKIP_WAL_OR_ROWS_REMOVED).
277+
*
278+
* Since, temporary slots are dropped after server restart, persisting
279+
* slotsync_skip_reason provides no practical benefit.
280+
*/
281+
SlotSyncSkipReasonslotsync_skip_reason;
252282
}ReplicationSlot;
253283

254284
#defineSlotIsPhysical(slot) ((slot)->data.database == InvalidOid)

0 commit comments

Comments
 (0)
[8]ページ先頭
©2009-2025 Movatter.jp