Re: "Stuck" query

Richard Huxton wrote:
> Tommy Gildseth wrote:
>> SELECT pg_cancel_backend(17504) has no effect, neither does kill 17504
>> from the shell.
>
> Strange.
>
>> I tried "strace -p17504", and this gave me just the following output:
>> sendto(7, "\7\0\0\0\003771\0\0\0\00224\0\0\0\017127.120.213.18"..., 968,
>> 0, NULL, 0 <unfinished ...>
>
> Looks like part of your query results being sent. Is it hung in that one
> system-call?


Yes, I left it there for about ~1 hour, and that was all that ever came.


>
>> Does anyone have any further troubleshooting suggestions that I can do,
>> to figure out why this query have "crashed"?
>
> Is the client locked/crashed too?
> If it's connected over a network, is the connection still there?


We stopped the client application to see if that would make any
difference, but the connection and the query on the server side is still
there.

A strace of the client application before it was shut down simply showed:

strace -p6721
Process 6721 attached - interrupt to quit
accept(18,  <unfinished ...>

so, not very helpfull.


>
>> The pg version is 8.2.9 on RHEL4
>
> Fairly standard setup. I've seen various problems reported by selinux
> oddities, but nothing quite like this.

We don't use selinux.


--
Tommy Gildseth