<div dir="ltr">Hi Ben, <div><br></div><div>Thanks for the help. The workload is a typical ETL processing. Building a DataWarehouse. I'll get you the rest of that information as soon as I can the tip about the metadata is already very interesting.</div>
<div><br></div><div>Thanks,</div><div>Erik</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Fri, Feb 22, 2013 at 9:25 PM, Ben Myers <span dir="ltr"><<a href="mailto:bpm@sgi.com" target="_blank">bpm@sgi.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi Erik,<br>
<div><div class="h5"><br>
On Fri, Feb 22, 2013 at 02:04:17PM +0100, Erik Knight wrote:<br>
> We've recently noticed that our system is experiencing extreme performance<br>
> problems when running large workloads. The problem seems to come from<br>
> excessive System CPU time. Specifically dozens of xfsaild threads. We used<br>
> to have SSD drives but recently switched to HDD, so some of us are thinking<br>
> that there may be a configuration issue within XFS that is optimized for<br>
> SSD but performs terribly slow on HDD.<br>
><br>
> Can anyone explain what these threads do, what would cause so many of them<br>
> to be running simultaneously or consume so much CPU?<br>
<br>
</div></div>AIL stands for Active Item List. These guys sync metadata which has been<br>
logged to its final location on disk. You'll have one daemon per filesystem.<br>
<br>
If you have a very metadata intensive workload they could get a workout. It<br>
would help to know a bit more about your workload and configuration. Can you<br>
provide the relevant information listed here?<br>
<br>
<a href="http://www.xfs.org/index.php/XFS_FAQ#Q:_What_information_should_I_include_when_reporting_a_problem.3F" target="_blank">http://www.xfs.org/index.php/XFS_FAQ#Q:_What_information_should_I_include_when_reporting_a_problem.3F</a><br>
<br>
Regards,<br>
Ben<br>
</blockquote></div><br></div>