No, haven't done any memory tests lately. Let me get that going.
I am doing pg_dump nightly with the whole database including this
table. so i guess that says it. I couldn't find any core dump. btw,
the core dump filename should be just like 'coredump', right?
Let me also get the max_stack_depth into conf.
Thanks much!!
telin
Thanks much for your advise.
Tom Lane wrote:
>"Telin Lin" <tlin@rosettastone.com> writes:
>
>
>>My postgresql keeps crashing by signal 11 after a number of concurrent
>>inserts + updates.
>>
>>
>
>This could be due to corrupt data, or it could be a hardware problem ---
>have you done any memory tests on that machine lately? If you can
>pg_dump the table involved without any failure, then the corrupt-data
>hypothesis is probably ruled out. Also, try to get a stack trace from
>the core dump (if no core file, restart postmaster under ulimit -c
>unlimited --- it's easiest to put that into the postgres init script).
>
> regards, tom lane
>
>