<div>Ok, I think I remember this discussion:</div><div><a href="http://gluster.org/pipermail/gluster-users/2012-February/009590.html">http://gluster.org/pipermail/gluster-users/2012-February/009590.html</a></div><div><br></div>
<div>Is this patch applied in future versions? Is it in general safe for me to use?</div><div><br></div><div><br></div><div><div class="gmail_quote">On Thu, Mar 1, 2012 at 9:09 AM, Anand Avati <span dir="ltr"><<a href="mailto:anand.avati@gmail.com">anand.avati@gmail.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="HOEnZb"><div class="h5">On Fri, Feb 24, 2012 at 3:23 PM, Bryan Whitehead <<a href="mailto:driver@megahappy.net">driver@megahappy.net</a>> wrote:<br>
> How long does it take for gluster volume set performance.io-thread-count 64<br>
> (as an example) to propagate?<br>
><br>
> I've noticed that it seems like mounted volumes don't get the performance<br>
> boost until i restart glusterd on the boxes. Is this wrong? Should I just be<br>
> more patient?<br>
><br>
> (note: I'm using infiniband and raid-10; "set performance.io-thread-count<br>
> 64" makes writes go from ~70MB/sec to ~200MB/sec).<br>
><br>
<br>
<br>
</div></div>the io-thread-count parameter is a configuration of the maximum number<br>
of threads (not the fixed number of threads). The actual number of<br>
threads scale up with the number of active (sys)calls in transit<br>
(today logarithmically, but we're making it more effective, closer to<br>
linear) with io-thread-count acting as a cap.<br>
<br>
Thanks,<br>
Avati<br>
</blockquote></div><br></div>