On Wed, Jun 16, 2010 at 1:15 PM, Craig Ringer
<craig@postnewspapers.com.au> wrote:
> On 16/06/10 16:56, Allan Kamau wrote:
>
>> The function I have mentioned above seems to "hang" indefinitely (now
>> 12hours and counting) while it took only 2secs for the many times it
>> was previously invoked from the client application (with slightly
>> different parameter value at each such time).
>
> Is there anything in pg_catalog.pg_locks that relates to that query?
>
> If you attach a debugger to the stalled backend and get a backtrace,
> what's the output? See:
>
> http://wiki.postgresql.org/wiki/Getting_a_stack_trace_of_a_running_PostgreSQL_backend_on_Linux/BSD
>
> If you strace the problem backend, does it appear to be doing anything?
>
> --
> Craig Ringer
>
> Tech-related writing: http://soapyfrogs.blogspot.com/
>
Thank you Craig, this could be what I am looking for, I am following
the HOWTO in the resource you've sent and will attempt to capture the
trace when the next time the backend "sleeps".
Allan.