Re: Wait event that should be reported while waiting for WAL archiving to finish

[Fujii Masao]

On 2020/02/13 12:28, Michael Paquier wrote:

On Thu, Feb 13, 2020 at 02:29:20AM +0900, Fujii Masao wrote:

When I saw pg_stat_activity.wait_event while pg_basebackup -X none
is waiting for WAL archiving to finish, it was either NULL or
CheckpointDone. I think this is confusing. What about introducing
new wait_event like WAIT_EVENT_BACKUP_WAIT_WAL_ARCHIVE
(BackupWaitWalArchive) and reporting it during that period?

Sounds like a good idea to me.  You need to be careful that this does
not overwrite more low-level wait event registration though, so that
could be more tricky than it looks at first sight.

Thanks for the advise! Patch attached.

I found that the wait events "LogicalRewriteTruncate" and
"GSSOpenServer" are not documented. I'm thinking to add
them into doc separately if ok.

Regards,

--
Fujii Masao
NTT DATA CORPORATION
Advanced Platform Technology Group
Research and Development Headquarters

diff --git a/doc/src/sgml/monitoring.sgml b/doc/src/sgml/monitoring.sgml
index 9129f79bbf..1e2f81d994 100644
--- a/doc/src/sgml/monitoring.sgml
+++ b/doc/src/sgml/monitoring.sgml
@@ -1479,6 +1479,10 @@ postgres   27093  0.0  0.0  30096  2752 ?        Ss   
11:34   0:00 postgres: ser
          <entry><literal>SyncRep</literal></entry>
          <entry>Waiting for confirmation from remote server during synchronous 
replication.</entry>
         </row>
+        <row>
+         <entry><literal>BackupWaitWalArchive</literal></entry>
+         <entry>Waiting for WAL files required for the backup to be 
successfully archived.</entry>
+        </row>
         <row>
          <entry morerows="2"><literal>Timeout</literal></entry>
          <entry><literal>BaseBackupThrottle</literal></entry>
diff --git a/src/backend/access/transam/xlog.c 
b/src/backend/access/transam/xlog.c
index 3813eadfb4..d79417b443 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -11095,6 +11095,7 @@ do_pg_stop_backup(char *labelfile, bool waitforarchive, 
TimeLineID *stoptli_p)
                seconds_before_warning = 60;
                waits = 0;
 
+               pgstat_report_wait_start(WAIT_EVENT_BACKUP_WAIT_WAL_ARCHIVE);
                while (XLogArchiveIsBusy(lastxlogfilename) ||
                           XLogArchiveIsBusy(histfilename))
                {
@@ -11120,6 +11121,7 @@ do_pg_stop_backup(char *labelfile, bool waitforarchive, 
TimeLineID *stoptli_p)
                                                                 "but the 
database backup will not be usable without all the WAL segments.")));
                        }
                }
+               pgstat_report_wait_end();
 
                ereport(NOTICE,
                                (errmsg("all required WAL segments have been 
archived")));
diff --git a/src/backend/postmaster/pgstat.c b/src/backend/postmaster/pgstat.c
index 7169509a79..6114ab44ee 100644
--- a/src/backend/postmaster/pgstat.c
+++ b/src/backend/postmaster/pgstat.c
@@ -3848,6 +3848,9 @@ pgstat_get_wait_ipc(WaitEventIPC w)
                case WAIT_EVENT_SYNC_REP:
                        event_name = "SyncRep";
                        break;
+               case WAIT_EVENT_BACKUP_WAIT_WAL_ARCHIVE:
+                       event_name = "BackupWaitWalArchive";
+                       break;
                        /* no default case, so that compiler will warn */
        }
 
diff --git a/src/include/pgstat.h b/src/include/pgstat.h
index aecb6013f0..7c40ebc3dc 100644
--- a/src/include/pgstat.h
+++ b/src/include/pgstat.h
@@ -853,7 +853,8 @@ typedef enum
        WAIT_EVENT_REPLICATION_ORIGIN_DROP,
        WAIT_EVENT_REPLICATION_SLOT_DROP,
        WAIT_EVENT_SAFE_SNAPSHOT,
-       WAIT_EVENT_SYNC_REP
+       WAIT_EVENT_SYNC_REP,
+       WAIT_EVENT_BACKUP_WAIT_WAL_ARCHIVE
 } WaitEventIPC;
 
 /* ----------