Sorry, I should have included the fact that PG was compiled from source on the same machine where it runs using gcc
4.1.0,and config.log has:
./configure --with-python --with-gssapi --enable-thread-safety
Bob Lunney
--- On Mon, 3/7/11, Bob Lunney <bob_lunney@yahoo.com> wrote:
> From: Bob Lunney <bob_lunney@yahoo.com>
> Subject: [ADMIN] PG Server Crash
> To: pgsql-admin@postgresql.org
> Date: Monday, March 7, 2011, 11:24 PM
> OK, I got a nasty surprise
> today. The server threw everybody out of the pool and
> logged this:
>
> 10.224.12.4(52763) idle: pthread_mutex_lock.c:80:
> __pthread_mutex_lock: Assertion `mutex->__data.__owner ==
> 0' failed.
> :4d45a006.78ec:2011-03-07 11:57:45.316 EST:LOG:
> server process (PID 21298) was terminated by signal 6:
> Aborted
> :4d45a006.78ec:2011-03-07 11:57:45.316 EST:LOG:
> terminating any other active server processes
>
> It then stopped all the backends, restarted, recovered and
> worked fine for the rest of the day.
>
> The particulars:
>
> SLES 10
> PG 8.4.1
> 8-way Intel Xeon E5345 @ 2.33GHz
> 32 GB RAM
>
> max_connections = 1500
> shared_buffers = 1536MB
> work_mem = 32MB
>
> maintenance_work_mem = 256MB
>
> checkpoint_segments = 2048
> effective_cache_size = 20GB
> effective_io_concurrency = 6
> random_page_cost = 2.0
>
>
> This server has been rock solid running 8.4.1 for over a
> year. It has Slony replicating to another mirror image
> server that has not had any problems whatsoever.
>
> Any ideas?
>
> Thanks in advance for all your help, and especially for the
> amazing database software!
>
> Regards,
>
> Bob Lunney
>
>
>
>
> --
> Sent via pgsql-admin mailing list (pgsql-admin@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-admin
>