Обсуждение: Sudden stop of postmaster - Looking for help

Hello,

My RH 7.3 PGSQL 7.2 suddenly died last night and left me with no idea
of what happened. I am looking for a logfile and not finding one with
anything interesting in it. Any clues as to where the postmaster
might've dumped a record of what happened?

Here's the message returned from pg_ctl with the -s option which
helped me to re-start the stubborn postmaster. Could someone help me
deduce whether this is a serious issue or not please.

TIA!!!

bash-2.05a$ pg_ctl start -D /var/lib/pgsql/data -s
pg_ctl: Another postmaster may be running.  Trying to start postmaster
anyway.
Found a pre-existing shared memory block (ID 196608) still in use.
If you're sure there are no old backends still running,
remove the shared memory block with ipcrm(1), or just
delete "/var/lib/pgsql/data/postmaster.pid".
pg_ctl: cannot start postmaster
Examine the log output.


bash-2.05a$ rm /var/lib/pgsql/data/postmaster.pid
bash-2.05a$ pg_ctl start -D /var/lib/pgsql/data -s
bash-2.05a$ DEBUG:  database system was interrupted at 2003-03-17
21:39:38 UTC
DEBUG:  checkpoint record is at 0/DC2A414
DEBUG:  redo record is at 0/DC2A414; undo record is at 0/0; shutdown
FALSE
DEBUG:  next transaction id: 1081353; next oid: 37710
DEBUG:  database system was not properly shut down; automatic recovery
in progress
DEBUG:  ReadRecord: record with zero length at 0/DC2A454
DEBUG:  redo is not required
DEBUG:  database system is ready


DAVE

Dave <not_here@due_to_spam.com> writes:
> bash-2.05a$ pg_ctl start -D /var/lib/pgsql/data -s
> pg_ctl: Another postmaster may be running.  Trying to start postmaster
> anyway.
> Found a pre-existing shared memory block (ID 196608) still in use.
> If you're sure there are no old backends still running,
> remove the shared memory block with ipcrm(1), or just
> delete "/var/lib/pgsql/data/postmaster.pid".

> bash-2.05a$ rm /var/lib/pgsql/data/postmaster.pid

Are you certain there wasn't a postmaster still running?  I think you
may have just shot yourself in the foot.  I'd recommend shutting down
postgres and making sure (with ps) that there are no remaining
postgres-owned processes.

            regards, tom lane

> Dave <not_here@due_to_spam.com> writes:
> > bash-2.05a$ pg_ctl start -D /var/lib/pgsql/data -s
> > pg_ctl: Another postmaster may be running.  Trying to start
> postmaster> anyway.
> > Found a pre-existing shared memory block (ID 196608) still in use.
> > If you're sure there are no old backends still running,
> > remove the shared memory block with ipcrm(1), or just
> > delete "/var/lib/pgsql/data/postmaster.pid".
>
> > bash-2.05a$ rm /var/lib/pgsql/data/postmaster.pid
>
> Are you certain there wasn't a postmaster still running?  I think you
> may have just shot yourself in the foot.  I'd recommend shutting down
> postgres and making sure (with ps) that there are no remaining
> postgres-owned processes.

I have postgres 7.3 (compiled from source code) in an IBM SMP XServer
with RH 8.0. And after beeing running 80 days, a partner called me
saying that our application cannot access to the database. Then I
logged in, and browse the processes (ps -ax), and there was not any
postmaster running. I looked the server.log 10 last lines, and seems as
an optimization, vacuum, statistics or something similar was executed.

I ran the postmaster again, and everything was fine. After it, I think
that an reboot could be better (so, It will relase memory), and I did
it.

¿ Where would be the problem ? I think, that in this case, pheraps, the
problem  could be similar than Dave's.

PD: This server is used only for serving datas with postgresql.