Re: BUG #6275: Horrible performance regression

Hi Mark

Thanks for your suggestions (and Robert and Guillaume). I will try the
overcommit_memory=2 idea out in a couple of weeks. At the moment I am
running everything on 8.4.8, which is working great, and have pressing
deadlines looming (next Wednesday).

I'll let you know when I do. I want to find a test case that exhibits a
real postgresql bug, or informs me as to what I have done wrong.

Regards,
Finaly

On Tue, 2011-11-01 at 17:26 +1300, Mark Kirkwood wrote:
> On 28/10/11 15:42, Finlay Thompson wrote:
> > After upgrading the postgresql*-8.4 packages on ubuntu, to version 8.4.9,
> > the script suddenly stopped working, and consuming all the ram (16GB) on the
> > computer (i7).
> >
> >
>
> If there is one query exhausting all ram, then it could be tricky to
> catch it in the logs (log duration and friends log the query time on
> completion... which might be never). If you have vm.overcommit_memory
> set to 2 then you may get the offending statement in the log when the
> postgres backend errors out with ENOMEM.
>
> If you know the approx time at which memory starts filling up then
> getting (most of) the query text from pg_stat_activity would enable you
> to track down the offending sql in your load script.
>
> regards
>
> Mark

--
Finlay Thompson
www.dragonfly.co.nz
Dragonfly, PO Box 27535, Wellington 6141, New Zealand
Level 5, 158 Victoria Street, Te Aro, Wellington
M: +64 21 426 878
T: +64 4 385 9285