512M is still REALLY high for a 32 bit postgresql. Have you tried something in the 16Meg range? On Tue, Aug 24, 2010 at 10:31 PM, Jeremy Palmer <JPalmer@xxxxxxxxxxxx> wrote: > Bugger I got another crash on the server today even after setting the temp_buffers to 512MB. Has anyone got any suggestions to fix this issue? > > Should I just compile the source using MS visual studio, then debug and get a stack trace for someone to diagnose on this list? > > Thanks > Jeremy > > -----Original Message----- > From: Jeremy Palmer > Sent: Friday, August 20, 2010 1:52 PM > To: 'Magnus Hagander'; Tom Lane > Cc: pgsql-general@xxxxxxxxxxxxxx; Alvaro Herrera; Chris Crook > Subject: RE: Win32 Backend Cash - pre-existing shared memory block is still in use > > Yes I do realise that temp_buffers is per backend. I set it like this because we only have a few simultaneous clients connecting, and these clients generally run large analysis queries that usually create big temp tables. > > I turned on extra logging and I have tracked down the query that is crashing the backend. The query was making a really big temp table. By setting the temp_buffers to 512MB the queries no longer crashes the backend. > > My question is what is a safe value for the temp_buffers parameter on a win32 system? Also how can we stop PostgreSQL crashing because of this issue? I'm willing provide more information to help diagnose this. > > Regards, > Jeremy > > -----Original Message----- > From: Magnus Hagander [mailto:magnus@xxxxxxxxxxxx] > Sent: Friday, August 20, 2010 1:46 AM > To: Tom Lane > Cc: Jeremy Palmer; pgsql-general@xxxxxxxxxxxxxx; Alvaro Herrera > Subject: Re: Win32 Backend Cash - pre-existing shared memory block is still in use > > On Thu, Aug 19, 2010 at 15:42, Tom Lane <tgl@xxxxxxxxxxxxx> wrote: >> Jeremy Palmer <JPalmer@xxxxxxxxxxxx> writes: >>> Could it be that I have too much memory allocated for postgresql? My resource settings are: >>> shared_buffers = 94952 >>> temp_buffers = 1GB >>> work_mem = 19339 >>> maintenance_work_mem = 191845 >>> max_stack_depth = 2MB >> >> 1GB for temp_buffers is a *LOT*. You do realize that's per backend? >> Those other settings don't look too unreasonable. > > Definitely - particularly since this is a 32-bit version, that's > getting very close to the address space limits... > > -- > Magnus Hagander > Me: http://www.hagander.net/ > Work: http://www.redpill-linpro.com/ > ______________________________________________________________________________________________________ > > This message contains information, which is confidential and may be subject to legal privilege. > If you are not the intended recipient, you must not peruse, use, disseminate, distribute or copy this message. > If you have received this message in error, please notify us immediately (Phone 0800 665 463 or info@xxxxxxxxxxxx) and destroy the original message. > LINZ accepts no responsibility for changes to this email, or for any attachments, after its transmission from LINZ. > > Thank you. > ______________________________________________________________________________________________________ > > -- > Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) > To make changes to your subscription: > http://www.postgresql.org/mailpref/pgsql-general > -- To understand recursion, one must first understand recursion. -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
