pg.ddx.io  pgsql-performance@postgresql.org mailing list archive  
help / color / mirror / Atom feed
From: Mark Kirkwood <markir@coretech.co.nz>
To: Stef <svb@ucs.co.za>
Cc: pgsql-performance@postgresql.org
Subject: Re: Strange problems with more memory.
Date: Tue, 17 Aug 2004 18:26:15 +1200
Message-ID: <4121A507.1000704@coretech.co.nz> (raw)
In-Reply-To: <20040816182314.24818138@svb.ucs.co.za>
References: <20040816171234.5a9bac8f@svb.ucs.co.za>
	<14009.1092670876@sss.pgh.pa.us>
	<20040816182314.24818138@svb.ucs.co.za>

If your nightly process is heavily read-only, then raid5 is probably 
fine. If however, there is a significant write component then it would 
perhaps be worth getting another disk and converting to raid10 
(alternatively - see previous postings about raid cards with on-board 
cache). Are you seeing a lot of write activity?

Note that it is possible for a SELECT only workload to generate 
significant write activity - if the resulting datasets are too large for 
memory sorting or hashing. I'm *guessing* that with an 11G database and 
1G (or was that 2G?) of ram that it is possible to overflow whatever 
your sort_mem is set to.


cheers

Mark

Stef wrote:

>Got 3 10000 rpm SCSI raid5 on here. I doubt I will get much better than that
>without losing both arms and legs... 
>
>  
>



view thread (4+ messages)

Message-ID: <4121A507.1000704@coretech.co.nz>
Permalink:  ../4121A507.1000704@coretech.co.nz/
Also on:    postgresql.org/message-id/4121A507.1000704@coretech.co.nz

 · 

reply

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Reply to all the recipients using the --to and --cc options:
  reply via email

  To: pgsql-performance@postgresql.org
  Cc: markir@coretech.co.nz, svb@ucs.co.za
  Subject: Re: Strange problems with more memory.
  In-Reply-To: <4121A507.1000704@coretech.co.nz>

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

This inbox is served by DDX for PostgreSQL; see mirroring instructions
for how to clone and mirror all data and code used for this inbox