Re: memcached and PostgreSQL

From: Sean Chittenden <sean(at)chittenden(dot)org>
To: Michael Adler <adler(at)pobox(dot)com>
Cc: pgsql-performance(at)postgresql(dot)org
Subject: Re: memcached and PostgreSQL
Date: 2004-11-18 21:10:28
Message-ID: 45A941C7-39A6-11D9-B82A-000A95C705DC@chittenden.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-performance

> So What does memcached offer pgsql users? It would still seem to offer
> the benefit of a multi-machined cache.

Ack, I totally missed this thread. Sorry for jumping in late.

Basically, memcached and pgmemcache offer a more technically correct
way of implementing query caching. MySQL's query caching is a
disaster, IMHO. memcached alleviates this load from the database and
puts it elsewhere in a more optimized form. The problem with memcached
by itself is that you're relying on the application to invalidate the
cache. How many different places have to be kept in sync? Using
memcached, in its current form, makes relying on the application to be
developed correctly with centralized libraries and database access
routines. Bah, that's a cluster f#$@ waiting to happen.

pgmemcache fixes that though so that you don't have to worry about
invalidating the cache in every application/routine. Instead you just
centralize that logic in the database and automatically invalidate via
triggers. It's working out very well for me.

I'd be interested in success stories, fwiw. In the next week or so
I'll probably stick this on pgfoundry and build a proper make/release
structure. -sc

--
Sean Chittenden

In response to

Browse pgsql-performance by date

  From Date Subject
Next Message Dawid Kuroczko 2004-11-19 13:59:48 When to bump up statistics?
Previous Message Dave Cramer 2004-11-18 19:46:41 Re: Interaction between Free Space Map an alternate location