Re: Bug#67481: Should a package increase a kernel parameter? (Re: Bug#67481)
On Thu, Jul 20, 2000 at 09:07:50PM +1200, Andrew McMillan wrote:
> Oliver Elphick wrote:
> >
> >
> > Example:
> > postgresql is configured to run 4096 backends (good grief!) and so
>
> Nope. 4096 _shared_buffers_ - quite a difference. I only have it set
> for 256 backends :-)
>
> Whacking up the buffers to use the memory available is a pretty standard
> heuristic to increase performance. Over time those buffers will pretty
> much cache the database, if the queries tend to hit the same records.
It's just worth pointing out here that the linux kernel has its own
buffers. If your machine has enough memory to cache the database
files, then the kernel will do it automagically, if they are are being
hit constantly.
It would take a postgres guru or an extended test to tell you what the
performance difference was between lots of SHMEM buffers and the linux
kernel auto-buffering, but I'd be surprised if there was much in it
speed-wise. (The former would be more dependable --- if something
else on the system reads another big file, your kernel buffers will be
displaced).
>
> We have a few machines here which have over 300M of RAM (not in the
> least unusual) and are starting to move towards ordering 256M as
> standard on newer ones. I can see 1GB coming soon enough (for us, I
> mean - lot's of people already have that much RAM).
>
Yes, that's certainly true.
Jules
Reply to: