[Spacewalk-list] Postgresql slowness

Jonathan Hoser jonathan.hoser at helmholtz-muenchen.de
Thu Jul 21 14:49:56 UTC 2011


On 07/21/2011 02:02 PM, John Hodrien wrote:
> On Thu, 21 Jul 2011, Jonathan Hoser wrote:
>> My config lines:
>> [...]
>> maintainance_work_mem = 32GB
>> fsync = off
>> synchronous_commit = off
>> checkpoint_segments = 30
>>
>>
>> I have not set 'wal_buffers' and 'effective_cache_size' so I guess these
>> are at default levels.
>> Files reside on a software (LVM) Raid1 consisting of two 1TB SATA drives.
> So you've got a beefy machine for postgres, and you don't particularly care
> about corruption of the database?
>
> I thought fsync = off was for people who don't really want a database all that
> much?
>
> jh
Well, I reread the passage from the manual, and there they state, that
as long as your hardware doesn't crash, you're fine.
Ah. Btw. maintainance_work_mem is only 32MB not GB. But 12 GB of shared
bufferes is - in my opinion - a requirement for medium databases;

So basically, since I'll lose at most ~23h in the database (nightly
backup), and not *that* much is going on, AND I am running it on a
system that I am confident will remain stable,
I can life with these settings.

Best
-Jonathan

Helmholtz Zentrum München
Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH)
Ingolstädter Landstr. 1
85764 Neuherberg
www.helmholtz-muenchen.de
Aufsichtsratsvorsitzende: MinDir´in Bärbel Brumme-Bothe
Geschäftsführer: Prof. Dr. Günther Wess und Dr. Nikolaus Blum
Registergericht: Amtsgericht München HRB 6466
USt-IdNr: DE 129521671




More information about the Spacewalk-list mailing list