Re: more problems with count(*) on large table

On Mon, Oct 01, 2007 at 01:34:32PM -0400, Bill Moran wrote:
> This sounds like a caching issue.  My guess at what's happening is that
> other operations are pushing this data out of the shared_buffers, so
> when you run it, the system has to pull a bunch of tuples off the disk
> to check them.  If you run it again immediately, the tuples are still in
> memory, and it runs very fast.

You should check your ratio of system cache to shared_buffers. Pushing
things out of shared buffers isn't bad if you've got a much bigger OS
behind it. The cost of pulling something out of the OS cache is
negligable compared to really going to disk.

Have a nice day,
-- 
Martijn van Oosterhout      http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.