Add wait events for recovery conflicts.

This commit introduces new wait events RecoveryConflictSnapshot and
RecoveryConflictTablespace. The former is reported while waiting for
recovery conflict resolution on a vacuum cleanup. The latter is reported
while waiting for recovery conflict resolution on dropping tablespace.

Also this commit changes the code so that the wait event Lock is reported
while waiting in ResolveRecoveryConflictWithVirtualXIDs() for recovery
conflict resolution on a lock. Basically the wait event Lock is reported
during that wait, but previously was not reported only when that wait
happened in ResolveRecoveryConflictWithVirtualXIDs().

Author: Masahiko Sawada
Reviewed-by: Fujii Masao
Discussion: https://postgr.es/m/CA+fd4k4mXWTwfQLS3RPwGr4xnfAEs1ysFfgYHvmmoUgv6Zxvmg@mail.gmail.com

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 0ebadf0d26d232835d4aa8e92e5e0f522c51ac1b..fd8b17ef8f6c38719a8e3bc0a8bb3dfcad9fbe9f 100644 (file)
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -1346,7 +1346,7 @@ postgres   27093  0.0  0.0  30096  2752 ?        Ss   11:34   0:00 postgres: ser
          <entry>Waiting in an extension.</entry>
         </row>
         <row>
-         <entry morerows="38"><literal>IPC</literal></entry>
+         <entry morerows="40"><literal>IPC</literal></entry>
          <entry><literal>BackupWaitWalArchive</literal></entry>
          <entry>Waiting for WAL files required for the backup to be successfully archived.</entry>
         </row>
@@ -1482,6 +1482,14 @@ postgres   27093  0.0  0.0  30096  2752 ?        Ss   11:34   0:00 postgres: ser
          <entry><literal>Promote</literal></entry>
          <entry>Waiting for standby promotion.</entry>
         </row>
+        <row>
+         <entry><literal>RecoveryConflictSnapshot</literal></entry>
+         <entry>Waiting for recovery conflict resolution on a vacuum cleanup.</entry>
+        </row>
+        <row>
+         <entry><literal>RecoveryConflictTablespace</literal></entry>
+         <entry>Waiting for recovery conflict resolution on dropping tablespace.</entry>
+        </row>
         <row>
          <entry><literal>RecoveryPause</literal></entry>
          <entry>Waiting for recovery to be resumed.</entry>

diff --git a/src/backend/postmaster/pgstat.c b/src/backend/postmaster/pgstat.c
index 04274056ca78ff5b274aaf9221e45617201f0b3e..9ebde47dea9f22eac474b0c59714751919dc734b 100644 (file)
--- a/src/backend/postmaster/pgstat.c
+++ b/src/backend/postmaster/pgstat.c
@@ -3852,6 +3852,12 @@ pgstat_get_wait_ipc(WaitEventIPC w)
        case WAIT_EVENT_PROMOTE:
            event_name = "Promote";
            break;
+       case WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT:
+           event_name = "RecoveryConflictSnapshot";
+           break;
+       case WAIT_EVENT_RECOVERY_CONFLICT_TABLESPACE:
+           event_name = "RecoveryConflictTablespace";
+           break;
        case WAIT_EVENT_RECOVERY_PAUSE:
            event_name = "RecoveryPause";
            break;

diff --git a/src/backend/storage/ipc/standby.c b/src/backend/storage/ipc/standby.c
index 08f695a98053002d6dd13d73a9162258fa582769..bdaf10a4b1f76cffe184ba66a2b334457b4aabb3 100644 (file)
--- a/src/backend/storage/ipc/standby.c
+++ b/src/backend/storage/ipc/standby.c
@@ -43,7 +43,9 @@ int           max_standby_streaming_delay = 30 * 1000;
 static HTAB *RecoveryLockLists;
 
 static void ResolveRecoveryConflictWithVirtualXIDs(VirtualTransactionId *waitlist,
-                                                  ProcSignalReason reason, bool report_waiting);
+                                                  ProcSignalReason reason,
+                                                  uint32 wait_event_info,
+                                                  bool report_waiting);
 static void SendRecoveryConflictWithBufferPin(ProcSignalReason reason);
 static XLogRecPtr LogCurrentRunningXacts(RunningTransactions CurrRunningXacts);
 static void LogAccessExclusiveLocks(int nlocks, xl_standby_lock *locks);
@@ -184,7 +186,7 @@ static int  standbyWait_us = STANDBY_INITIAL_WAIT_US;
  * more then we return true, if we can wait some more return false.
  */
 static bool
-WaitExceedsMaxStandbyDelay(void)
+WaitExceedsMaxStandbyDelay(uint32 wait_event_info)
 {
    TimestampTz ltime;
 
@@ -198,7 +200,9 @@ WaitExceedsMaxStandbyDelay(void)
    /*
     * Sleep a bit (this is essential to avoid busy-waiting).
     */
+   pgstat_report_wait_start(wait_event_info);
    pg_usleep(standbyWait_us);
+   pgstat_report_wait_end();
 
    /*
     * Progressively increase the sleep times, but not to more than 1s, since
@@ -223,7 +227,8 @@ WaitExceedsMaxStandbyDelay(void)
  */
 static void
 ResolveRecoveryConflictWithVirtualXIDs(VirtualTransactionId *waitlist,
-                                      ProcSignalReason reason, bool report_waiting)
+                                      ProcSignalReason reason, uint32 wait_event_info,
+                                      bool report_waiting)
 {
    TimestampTz waitStart = 0;
    char       *new_status;
@@ -264,7 +269,7 @@ ResolveRecoveryConflictWithVirtualXIDs(VirtualTransactionId *waitlist,
            }
 
            /* Is it time to kill it? */
-           if (WaitExceedsMaxStandbyDelay())
+           if (WaitExceedsMaxStandbyDelay(wait_event_info))
            {
                pid_t       pid;
 
@@ -317,6 +322,7 @@ ResolveRecoveryConflictWithSnapshot(TransactionId latestRemovedXid, RelFileNode
 
    ResolveRecoveryConflictWithVirtualXIDs(backends,
                                           PROCSIG_RECOVERY_CONFLICT_SNAPSHOT,
+                                          WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT,
                                           true);
 }
 
@@ -346,6 +352,7 @@ ResolveRecoveryConflictWithTablespace(Oid tsid)
                                                InvalidOid);
    ResolveRecoveryConflictWithVirtualXIDs(temp_file_users,
                                           PROCSIG_RECOVERY_CONFLICT_TABLESPACE,
+                                          WAIT_EVENT_RECOVERY_CONFLICT_TABLESPACE,
                                           true);
 }
 
@@ -417,6 +424,7 @@ ResolveRecoveryConflictWithLock(LOCKTAG locktag)
         */
        ResolveRecoveryConflictWithVirtualXIDs(backends,
                                               PROCSIG_RECOVERY_CONFLICT_LOCK,
+                                              PG_WAIT_LOCK | locktag.locktag_type,
                                               false);
    }
    else

diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index 9d351e771452e3abcd344e68d0287a4f50a0c950..b8041d9988046554c6fff4192b5bf1cdbce8f79f 100644 (file)
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -881,6 +881,8 @@ typedef enum
    WAIT_EVENT_PARALLEL_FINISH,
    WAIT_EVENT_PROCARRAY_GROUP_UPDATE,
    WAIT_EVENT_PROMOTE,
+   WAIT_EVENT_RECOVERY_CONFLICT_SNAPSHOT,
+   WAIT_EVENT_RECOVERY_CONFLICT_TABLESPACE,
    WAIT_EVENT_RECOVERY_PAUSE,
    WAIT_EVENT_REPLICATION_ORIGIN_DROP,
    WAIT_EVENT_REPLICATION_SLOT_DROP,