We're testing application with Postgres 7.3.2 on the backend.
The entire test involves running series of individual tests.
Few times throughout this procedure the database server is being
shut down and started up again
Normally, this is what I see in the database log.
2003-07-29 10:14:33 [14513] LOG: smart shutdown request
2003-07-29 10:14:33 [14594] LOG: shutting down
2003-07-29 10:14:35 [14594] LOG: database system is shut down
2003-07-29 10:14:37 [14676] LOG: database system was shut down at
2003-07-29 10:14:35 EDT
2003-07-29 10:14:37 [14676] LOG: checkpoint record is at 0/9383DC
2003-07-29 10:14:37 [14676] LOG: redo record is at 0/9383DC; undo record
is at 0/0; shutdown TRUE
2003-07-29 10:14:37 [14676] LOG: next transaction id: 1899; next oid:
17344
2003-07-29 10:14:37 [14676] LOG: database system is ready
Now, (on Linux only) we sometimes run into this:
2003-07-29 11:31:15 [26665] LOG: smart shutdown request
2003-07-29 11:31:15 [26728] LOG: shutting down
2003-07-29 11:31:17 [26728] LOG: database system is shut down
2003-07-29 11:31:19 [26767] LOG: database system was shut down at
2003-07-29 11:31:17 EDT
2003-07-29 11:31:19 [26767] LOG: checkpoint record is at 0/C01520
2003-07-29 11:31:19 [26767] LOG: redo record is at 0/C01520; undo record
is at 0/0; shutdown TRUE
2003-07-29 11:31:19 [26767] LOG: next transaction id: 15648; next oid:
18523
2003-07-29 11:31:24 [26769] FATAL: The database system is starting up
2003-07-29 11:31:26 [26767] LOG: database system is ready
The database was running with
LOG_MIN_ERROR_STATEMENT = DEBUG2
in postgresql.conf, but there were no additional information in the database
log.
I am wondering if on Linux it takes more time to shut down database server,
and we get this because we try to start the server while some of postgres
processes are still around?
Any ideas are appreciated.
Thanks,
Mike.