From: | Greg Smith <greg(at)2ndquadrant(dot)com> |
---|---|
To: | Ma Sivakumar <masivakumar(at)gmail(dot)com> |
Cc: | pgsql-general(at)postgresql(dot)org |
Subject: | Re: MySQL versus Postgres |
Date: | 2010-08-12 05:23:44 |
Message-ID: | 4C638560.2060400@2ndquadrant.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
Ma Sivakumar wrote:
> There can be a stronger and more prominent hint / recommendation in
> postgresql.conf file, in install README, Resource Consumption section
> of manual (http://www.postgresql.org/docs/8.4/interactive/runtime-config-resource.html)
> and other appropriate places
>
What you might not be aware of is that most of the text you quoted was
just added to the PostgreSQL documentation in the last couple of
months. That already represents a good step forward from earlier
versions; until quite recently
http://www.postgresql.org/docs/8.2/interactive/runtime-config-resource.html
was how that looked even for 8.3 and 8.4. I'm happy that we've already
made a big step forward in helping new users here in that update. At
least now DBAs used to other systems who go looking for the "how do I
set the size of the cache?" knob will stumble on reasonable advice in
the manual, which was not the case before. So I already expect this
situation to be much better than it has ever been.
> Can a final
> "We recommend that you to set shared_buffers to 25% of system memory
> you can allocate for PostgreSQL"
> be included.
>
If only it were that easy. 25%, but only on a dedicated server, don't
go above 8GB, limit to much less than that on Windows, and be extremely
careful if you're writing heavily lest large checkpoints squash you.
Giving simple advice that people might follow without thinking about
actually has some worst-case downsides that are worse than not tuning
the server at all.
> This makes Increasing work_mem is a complex calculation for a new user
> trying out PostgreSQL.
>
Increasing work_mem is a complex calculation for everyone, not just new
users. If it were easy for anyone, we'd just bottle whatever experts do
into the software directly. I tried doing a round of that with pgtune,
and it's still not quite right yet even after surveying everyone who had
an opinion on the subject.
--
Greg Smith 2ndQuadrant US Baltimore, MD
PostgreSQL Training, Services and Support
greg(at)2ndQuadrant(dot)com www.2ndQuadrant.us
From | Date | Subject | |
---|---|---|---|
Next Message | Sandeep Srinivasa | 2010-08-12 05:35:55 | Re: MySQL versus Postgres |
Previous Message | Ma Sivakumar | 2010-08-12 04:57:20 | Re: MySQL versus Postgres |