Re: pl/pgsql function spikes CPU 100%

Поиск
Список
Период
Сортировка
От Jeff Frost
Тема Re: pl/pgsql function spikes CPU 100%
Дата
Msg-id Pine.LNX.4.64.0703160827490.12217@discord.home.frostconsultingllc.com
обсуждение исходный текст
Ответ на Re: pl/pgsql function spikes CPU 100%  (Jeff Frost <jeff@frostconsultingllc.com>)
Ответы Re: pl/pgsql function spikes CPU 100%
Список pgsql-admin
On Fri, 16 Mar 2007, Jeff Frost wrote:

> On Fri, 16 Mar 2007, Tom Lane wrote:
>
>> Jeff Frost <jeff@frostconsultingllc.com> writes:
>>> ... Interestingly, when you
>>> strace the backend, it doesn't appear to be doing too much...here's some
>>> sample output:
>>
>>> select(0, NULL, NULL, NULL, {0, 1000})  = 0 (Timeout)
>>> semop(3932217, 0x7fbfffd150, 1)         = 0
>>> semop(3932217, 0x7fbfffd150, 1)         = 0
>>> semop(3932217, 0x7fbfffd150, 1)         = 0
>>> semop(3932217, 0x7fbfffd150, 1)         = 0
>>> semop(3932217, 0x7fbfffd150, 1)         = 0
>>> select(0, NULL, NULL, NULL, {0, 1000})  = 0 (Timeout)
>>> semop(3997755, 0x7fbfffd170, 1)         = 0
>>> semop(3932217, 0x7fbfffd150, 1)         = 0
>>
>> This looks suspiciously like the sort of trace we saw in the various
>> "context swap storm" threads.  The test cases for those generally
>> involved really tight indexscan loops, ie, the backends were spending
>> all their time trying to access shared buffers.  If you haven't changed
>> the function or the data, then I concur with the nearby worry about
>> autovacuuming (large buildup of dead tuples could result in this symptom).
>> Or maybe you've got an old open transaction that is blocking cleanup?
>
> Tom,
>
> I doubt it's a problem with autovacuum as the data in this server was just
> loaded an hour before the strace above was taken, so there should have been
> almost no dead tuples, especially since these tables are nearly write once.
> I.e. they get written to once, then the populate function updates them, then
> months later they get archived off.
>
> There did not appear to be high context switch activity nor any IO wait to
> mention during the time I was watching the postmaster.  If it's worth
> mentioning, it's running CentOS 4.4 with the kernel-2.6.9-34.EL kernel.

Oh, and I meant to mention that this query was the only thing in
pg_stat_activity during the painful time it was running, and there were no
ungranted locks in pg_locks.

--
Jeff Frost, Owner     <jeff@frostconsultingllc.com>
Frost Consulting, LLC     http://www.frostconsultingllc.com/
Phone: 650-780-7908    FAX: 650-649-1954

В списке pgsql-admin по дате отправления:

Предыдущее
От: Jeff Frost
Дата:
Сообщение: Re: pl/pgsql function spikes CPU 100%
Следующее
От: cedric
Дата:
Сообщение: unsubscribe