From c972f22a2697f54cda71b6b4e7b7f0eac477e9af Mon Sep 17 00:00:00 2001 From: Kyotaro Horiguchi Date: Thu, 21 Dec 2017 21:23:25 +0900 Subject: [PATCH 2/3] Add monitoring aid for max_replication_slots. Adds two columns "status" and "min_secure_lsn" in pg_replication_slot. Setting max_slot_wal_keep_size, long-disconnected slots may lose sync. The two columns shows that a slot can be reconnected or not, or about to lose required WAL segments. And the LSN back to where the next checkpoint will secure. --- src/backend/access/transam/xlog.c | 93 ++++++++++++++++++++++++++++++++++++ src/backend/catalog/system_views.sql | 4 +- src/backend/replication/slotfuncs.c | 25 +++++++++- src/include/access/xlog.h | 1 + src/include/catalog/pg_proc.h | 2 +- src/test/regress/expected/rules.out | 6 ++- 6 files changed, 126 insertions(+), 5 deletions(-) diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c index 723a983..b630224 100644 --- a/src/backend/access/transam/xlog.c +++ b/src/backend/access/transam/xlog.c @@ -9349,6 +9349,99 @@ CreateRestartPoint(int flags) return true; } + +/* + * Returns the segment number of the oldest file in XLOG directory. + */ +static XLogSegNo +GetOldestXLogFileSegNo(void) +{ + DIR *xldir; + struct dirent *xlde; + XLogSegNo segno = 0; + + xldir = AllocateDir(XLOGDIR); + if (xldir == NULL) + ereport(ERROR, + (errcode_for_file_access(), + errmsg("could not open write-ahead log directory \"%s\": %m", + XLOGDIR))); + + while ((xlde = ReadDir(xldir, XLOGDIR)) != NULL) + { + TimeLineID tli; + XLogSegNo fsegno; + + /* Ignore files that are not XLOG segments */ + if (!IsXLogFileName(xlde->d_name) && + !IsPartialXLogFileName(xlde->d_name)) + continue; + + XLogFromFileName(xlde->d_name, &tli, &fsegno, wal_segment_size); + + /* get minimum segment ignorig timeline ID */ + if (segno == 0 || fsegno < segno) + segno = fsegno; + } + + return segno; +} + +/* + * Check if the record on the given restartLSN is present in XLOG files. + * + * Returns true if it is present. If minSecureLSN is given, it receives the + * LSN at the beginning of the oldest existing WAL segment. + */ +bool +IsLsnStillAvaiable(XLogRecPtr restartLSN, XLogRecPtr *minSecureLSN) +{ + XLogRecPtr currpos; + XLogSegNo restartSeg; + XLogSegNo tailSeg; + XLogSegNo oldestSeg; + + Assert(!XLogRecPtrIsInvalid(restartLSN)); + + currpos = GetXLogWriteRecPtr(); + + SpinLockAcquire(&XLogCtl->info_lck); + oldestSeg = XLogCtl->lastRemovedSegNo; + SpinLockRelease(&XLogCtl->info_lck); + + /* + * oldestSeg is zero before at least one segment has been removed since + * startup. Use oldest segno taken from file names. + */ + if (oldestSeg == 0) + { + static XLogSegNo oldestFileSeg = 0; + + if (oldestFileSeg == 0) + oldestFileSeg = GetOldestXLogFileSegNo(); + /* let it have the same meaning with lastRemovedSegNo here */ + oldestSeg = oldestFileSeg - 1; + } + + /* oldest segment is just after the last removed segment */ + oldestSeg++; + + XLByteToSeg(restartLSN, restartSeg, wal_segment_size); + + + if (minSecureLSN) + { + XLogRecPtr slotPtr = XLogGetReplicationSlotMinimumLSN(); + Assert(!XLogRecPtrIsInvalid(slotPtr)); + + tailSeg = GetMinSecuredSegment(currpos, slotPtr); + + XLogSegNoOffsetToRecPtr(tailSeg, 0, *minSecureLSN, wal_segment_size); + } + + return oldestSeg <= restartSeg; +} + /* * Returns minimum segment number the next checktpoint must leave considering * wal_keep_segments, replication slots and max_slot_wal_keep_size. diff --git a/src/backend/catalog/system_views.sql b/src/backend/catalog/system_views.sql index 394aea8..4167146 100644 --- a/src/backend/catalog/system_views.sql +++ b/src/backend/catalog/system_views.sql @@ -793,7 +793,9 @@ CREATE VIEW pg_replication_slots AS L.xmin, L.catalog_xmin, L.restart_lsn, - L.confirmed_flush_lsn + L.confirmed_flush_lsn, + L.status, + L.min_secure_lsn 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 ab776e8..3880807 100644 --- a/src/backend/replication/slotfuncs.c +++ b/src/backend/replication/slotfuncs.c @@ -182,7 +182,7 @@ pg_drop_replication_slot(PG_FUNCTION_ARGS) Datum pg_get_replication_slots(PG_FUNCTION_ARGS) { -#define PG_GET_REPLICATION_SLOTS_COLS 11 +#define PG_GET_REPLICATION_SLOTS_COLS 13 ReturnSetInfo *rsinfo = (ReturnSetInfo *) fcinfo->resultinfo; TupleDesc tupdesc; Tuplestorestate *tupstore; @@ -304,6 +304,29 @@ pg_get_replication_slots(PG_FUNCTION_ARGS) else nulls[i++] = true; + if (restart_lsn == InvalidXLogRecPtr) + { + values[i++] = CStringGetTextDatum("unknown"); + values[i++] = LSNGetDatum(InvalidXLogRecPtr); + } + else + { + XLogRecPtr min_secure_lsn; + char *status = "broken"; + + if (BoolGetDatum(IsLsnStillAvaiable(restart_lsn, + &min_secure_lsn))) + { + if (min_secure_lsn <= restart_lsn) + status = "secured"; + else + status = "insecured"; + } + + values[i++] = CStringGetTextDatum(status); + values[i++] = LSNGetDatum(min_secure_lsn); + } + tuplestore_putvalues(tupstore, tupdesc, values, nulls); } LWLockRelease(ReplicationSlotControlLock); diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h index 45eb51a..542df28 100644 --- a/src/include/access/xlog.h +++ b/src/include/access/xlog.h @@ -269,6 +269,7 @@ extern void ShutdownXLOG(int code, Datum arg); extern void InitXLOGAccess(void); extern void CreateCheckPoint(int flags); extern bool CreateRestartPoint(int flags); +extern bool IsLsnStillAvaiable(XLogRecPtr restartLSN, XLogRecPtr *minSecureLSN); extern void XLogPutNextOid(Oid nextOid); extern XLogRecPtr XLogRestorePoint(const char *rpName); extern void UpdateFullPageWrites(void); diff --git a/src/include/catalog/pg_proc.h b/src/include/catalog/pg_proc.h index c969375..1157438 100644 --- a/src/include/catalog/pg_proc.h +++ b/src/include/catalog/pg_proc.h @@ -5340,7 +5340,7 @@ DATA(insert OID = 3779 ( pg_create_physical_replication_slot PGNSP PGUID 12 1 0 DESCR("create a physical replication slot"); DATA(insert OID = 3780 ( pg_drop_replication_slot PGNSP PGUID 12 1 0 0 0 f f f f t f v u 1 0 2278 "19" _null_ _null_ _null_ _null_ _null_ pg_drop_replication_slot _null_ _null_ _null_ )); DESCR("drop a replication slot"); -DATA(insert OID = 3781 ( pg_get_replication_slots PGNSP PGUID 12 1 10 0 0 f f f f f t s s 0 0 2249 "" "{19,19,25,26,16,16,23,28,28,3220,3220}" "{o,o,o,o,o,o,o,o,o,o,o}" "{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn}" _null_ _null_ pg_get_replication_slots _null_ _null_ _null_ )); +DATA(insert OID = 3781 ( pg_get_replication_slots PGNSP PGUID 12 1 10 0 0 f f f f f t s s 0 0 2249 "" "{19,19,25,26,16,16,23,28,28,3220,3220,25,3220}" "{o,o,o,o,o,o,o,o,o,o,o,o,o}" "{slot_name,plugin,slot_type,datoid,temporary,active,active_pid,xmin,catalog_xmin,restart_lsn,confirmed_flush_lsn,status,min_secure_lsn}" _null_ _null_ pg_get_replication_slots _null_ _null_ _null_ )); DESCR("information about replication slots currently in use"); DATA(insert OID = 3786 ( pg_create_logical_replication_slot PGNSP PGUID 12 1 0 0 0 f f f f t f v u 3 0 2249 "19 19 16" "{19,19,16,25,3220}" "{i,i,i,o,o}" "{slot_name,plugin,temporary,slot_name,lsn}" _null_ _null_ pg_create_logical_replication_slot _null_ _null_ _null_ )); DESCR("set up a logical replication slot"); diff --git a/src/test/regress/expected/rules.out b/src/test/regress/expected/rules.out index f1c1b44..d9d74a3 100644 --- a/src/test/regress/expected/rules.out +++ b/src/test/regress/expected/rules.out @@ -1451,8 +1451,10 @@ pg_replication_slots| SELECT l.slot_name, l.xmin, l.catalog_xmin, l.restart_lsn, - l.confirmed_flush_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) + l.confirmed_flush_lsn, + l.status, + l.min_secure_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, status, min_secure_lsn) LEFT JOIN pg_database d ON ((l.datoid = d.oid))); pg_roles| SELECT pg_authid.rolname, pg_authid.rolsuper, -- 2.9.2