Re: Dying PostgreSQL backend - Mailing list pgsql-admin

From otisg
Subject Re: Dying PostgreSQL backend
Date
Msg-id 038a01c1f623$d83191a0$6ec8010a@mail2world.com
Whole thread Raw
In response to Dying PostgreSQL backend  ("otisg" <otisg@iVillage.com>)
List pgsql-admin
Hello,

> What I'd recommend doing is gdb'ing the core dump file to get a stack
> trace; that would give us some clue what's wrong. And you could do
> "p debug_query_string" to see what query made it crash.
>
> If you do not find a core file in the database subdirectory
> ($PGDATA/base/yourdbnumber/core) then you are probably running the
> postmaster with "ulimit -c 0" which disables core dumps. Restart
> it with environment "ulimit -c unlimited".

Yes, I fixed that.

Unfortunately I can't gdb to look into the core file:
(gdb) p debug_query_string
No symbol table is loaded. Use the "file" command.
(gdb) file /tmp/core4
"/tmp/core4": not in executable format: File format not recognized
(gdb)

Any ideas?

I did, however, see this in the syslogd log at the time of core dump:

/usr/bin/postmaster child[4014]: starting with (postgres -d16 -v131072 -p mydb )
invoking IpcMemoryCreate(size=2015232)

The first line I had seen before (every few seconds, since my application makes 5-7 updates per second), but never the second one. This was the first time.
Maybe it means something to those familiar with the guts of PostgreSQL.

Thanks,
Otis
_______________________________________________________________
Sign up for FREE iVillage newsletters.
From health and pregnancy to shopping and relationships, iVillage
has the scoop on what matters most to you.

pgsql-admin by date:

Previous
From: Hal Lynch
Date:
Subject: batch operations
Next
From: Tom Lane
Date:
Subject: Re: db recovery (FATAL 2)