Skip site navigation (1) Skip section navigation (2)

Re: shared_buffers advice

From: Konrad Garus <konrad(dot)garus(at)gmail(dot)com>
To: Cédric Villemain <cedric(dot)villemain(dot)debian(at)gmail(dot)com>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: shared_buffers advice
Date: 2010-05-27 07:24:23
Message-ID: AANLkTikvDqFe6exbzGYhuQMw9NzIKbQ91qzTB-BOcDTT@mail.gmail.com (view raw or flat)
Thread:
Lists: pgsql-performance
2010/5/26 Cédric Villemain <cedric(dot)villemain(dot)debian(at)gmail(dot)com>:

> At the moment where a block is requested for the first time (usualy
> 8kb from postgres, so in fact 2 blocks in OS), you have 'duplicate'
> buffers.
> But, depending of your workload, it is not so bad because those 2
> blocks should not be requested untill some time (because in postgresql
> shared buffers) and should be evicted by OS in favor of new blocks
> requests.

Since pg_buffercache is 4-8 times smaller, it would seem to be
extremely rare to me. And when PG requests a block, it also needs to
evict something from shared_buffers.

> You can try pgfincore extension to grab stats from OS cache and/or
> patch postgresql if you want real stats ;)

Thank you! It seems to be the tool I was looking for. Could help me
locate and troubleshoot the hogs in page cache. I also find the
snapshot/restore function promising. Every morning our cache is cold
or filled with irrelevant data left by nightly batch jobs, thus
severely impacting the performance. Seems to be exactly what this tool
is for.

How does it work? How stable is it? Can we use it in production on a
daily basis?

> pgbuffercache is provided with postgresql and deliver very usefull information :
> http://www.postgresql.org/docs/8.4/interactive/pgbuffercache.html

Thank you. I already am using it. I've already found a few hogs with it.

-- 
Konrad Garus

In response to

Responses

pgsql-performance by date

Next:From: 黄永卫Date: 2010-05-27 07:27:41
Subject: About Tom Lane's Xeon CS test case
Previous:From: David JarvisDate: 2010-05-26 20:21:19
Subject: Re: Random Page Cost and Planner

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group