OT: Re: [exim] Reducing Disk IO - Speed Issues

Top Page
Delete this message
Reply to this message
Author: V. T. Mueller
Date:  
To: 'Exim-users'
Subject: OT: Re: [exim] Reducing Disk IO - Speed Issues
Hello,

Marc Perkel schrieb:
> I didn't know raid controllers had that feature. I'm using software raid
> but if I ever get a raid controller I'll look for that.
> Walt Reed wrote:
>> In addition to other suggestions that were put forth, making sure your
>> RAID controller has battery backed write cache does AMAZING things for
>> performance.
>> On Wed, Aug 24, 2005 at 08:15:36PM -0700, Marc Perkel said:
>>> I'm using EXT3 and wondering what tricks I should use to make disk IO
>>> faster and/or reduce disk IO. What are the tricks?


@Marc: a battery-backed write cache is an optional hardware feature
for the better controllers. This is somewhat paradox, because on
important high-IO systems - which is where you usually use rather
expensive RAID controllers - you rely completely upon UPS triggered
shutdown routines in case of a power failure as long as your
controller cache is not battery backed. It's quite tricky to restore
a database that had just lost a couple of megabytes of transactions...

@Walt: for the same reason it is always a good move to disable the
disk cache. Disabling asynchronous disk access might also be needed.
For a fixed amount of money, there is always a tradeoff between data
security/integrity and performance.

What has been suggested for the OS side - no-atime etc. - is the
more clever approach as long as you don't have software that relies
on it. If I remember right Nigel had put a couple of posts on the
list which summon this precisely.

Kind regards,
vt