On Fri, Mar 28, 2008 at 01:38:57PM -0500, Alex Adriaanse wrote: > First some background information: > > Software (at the time of the memory errors): CentOS 4.5 (x86_64) running > its 2.6.9-55.ELsmp Linux kernel, PostgreSQL 8.1.9 (from RPMs provided on > the PostgreSQL web site: postgresql-8.1.9-1PGDG.x86_64). > > Hardware: 4 dual-core Opterons. 16GB physical RAM, 2GB swap. Since you've clearly disabled overcommit, you should realise that under the default settings you will never use more than about 50-60% of your memory for the database. Your swap will essentially always be unused. The rest will be simply buffers/cache (which may be what you want). You may want to play with the VM settings and/or increase the swap. As for the precise cause of the crash: I don't see anything in the log that would indicate anything unusual, though evidently it wanted about 256MB all of a sudden, which looks suspicously like your maintainence workmem. My guess is that the total allocated VM is fairly close to the limit of your configuration and that a slightly higher than normal load and a background autovacuum took it over the edge. My guess is that if you up the swap to 4GB or perhaps 8GB, you'll get a lot more room to manouver. Does Zenoss track to *allocated* VM in addition to total memory used? Hope this helps, -- Martijn van Oosterhout <kleptog@xxxxxxxxx> http://svana.org/kleptog/ > Please line up in a tree and maintain the heap invariant while > boarding. Thank you for flying nlogn airlines.
Attachment:
signature.asc
Description: Digital signature
