Re: Scaling with memory & disk planning

From: Scott Marlowe <scott(dot)marlowe(at)ihs(dot)com>
To: <terry(at)greatgulfhomes(dot)com>
Cc: <pgsql-general(at)postgresql(dot)org>
Subject: Re: Scaling with memory & disk planning
Date: 2002-05-30 22:34:23
Message-ID: Pine.LNX.4.33.0205301622060.15639-100000@css120.ihs.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Thu, 30 May 2002 terry(at)greatgulfhomes(dot)com wrote:

> What is RAID1 with > 2 drives???

Just that. A mirror set that writes all changes to all drives.

It costs more for writes, but runs faster for reads, so it's not a good
idea to use one in a heavily written environment, but it really flies for
heavy read setups, like data mining or CRM stuff. Plus, a single drive
failure is no big deal, you just replace it at your leisure, since you're
still fully redundant.

This means you could do crazy things like build 9 disks into a RAID 0+1 by
making three sets of RAID 1 with three drives, and then put all three
together to make the RAID 0. Very good read speeds under heavy loads, and
fairly good write speeds as well.

You could run a fairly large database on a setup of 9 36 Gig ultra SCSI
drives like this, and get some nice throughput, and have it be VERY
redundant. I.e. ANY two drives could fail, and you'd still be up, and if
the two drives that failed were in different RAID1s, you'd still be fully
redundant.

Crazy stuff.

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Francisco Reyes 2002-05-30 23:35:03 Re: Moving data from FreeBSD to Red Hat
Previous Message terry 2002-05-30 22:12:00 Re: Scaling with memory & disk planning