<div dir="ltr">Hi Bryan,<div><br></div><div style>We have 1G nics on all our servers.</div><div style>Do you think that changing our design to distribute-replicate will improve the performance?</div><div style><br></div><div style>
Anything in the gluster performance settings that you think I should change?</div><div style><br></div><div style>thanks!</div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Tue, Apr 9, 2013 at 6:00 AM, Bryan Whitehead <span dir="ltr"><<a href="mailto:driver@megahappy.net" target="_blank">driver@megahappy.net</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">This looks like you are replicating every file to all bricks?<br>
<br>
What is tcp running on? 1G nics? 10G? IPoIB (40-80G)?<br>
<br>
I think you want to have Distribute-Replicate. So 4 bricks with replica = 2.<br>
<br>
Unless you are running at least 10G nics you are going to have serious<br>
IO issues in your KVM/qcow2 VM's.<br>
<div><div class="h5"><br>
On Mon, Apr 8, 2013 at 7:11 AM, Eyal Marantenboim<br>
<<a href="mailto:eyal@theserverteam.com">eyal@theserverteam.com</a>> wrote:<br>
> Hi,<br>
><br>
><br>
> We have a set of 4 gluster nodes, all in replicated (design?)<br>
><br>
> We use it to store our qcow2 images for kvm. These images have a variable<br>
> IO, though most of them are for reading only.<br>
><br>
><br>
> I tried to find some documentation re. performance optimization, but it's<br>
> either unclear to me, or I couldn't find much.. so I'd copied from the<br>
> internet and tried to adjust the config to our needs, but I'm sure it's not<br>
> optimized.<br>
><br>
><br>
> We're using 3.3.1 on top of XFS.<br>
><br>
> The qcow images are about 30GB (a couple of 100GB).<br>
><br>
><br>
> Can someone please tell what would be the best paramenter for performance to<br>
> look at?<br>
><br>
><br>
> here is volume info:<br>
><br>
><br>
> Volume Name: images<br>
><br>
> Type: Replicate<br>
><br>
> Volume ID:<br>
><br>
> Status: Started<br>
><br>
> Number of Bricks: 1 x 4 = 4<br>
><br>
> Transport-type: tcp<br>
><br>
> Bricks:<br>
><br>
> Brick1: vmhost2:/exports/1<br>
><br>
> Brick2: vmhost3:/exports/1<br>
><br>
> Brick3: vmhost5:/exports/1<br>
><br>
> Brick4: vmhost6:/exports/1<br>
><br>
> Options Reconfigured:<br>
><br>
> performance.cache-max-file-size: 1GB<br>
><br>
> nfs.disable: on<br>
><br>
> performance.cache-size: 4GB<br>
><br>
> performance.cache-refresh-timeout: 1<br>
><br>
> performance.write-behind-window-size: 2MB<br>
><br>
> performance.read-ahead: on<br>
><br>
> performance.write-behind: on<br>
><br>
> performance.io-cache: on<br>
><br>
> performance.stat-prefetch: on<br>
><br>
> performance.quick-read: on<br>
><br>
> performance.io-thread-count: 64<br>
><br>
> performance.flush-behind: on<br>
><br>
> features.quota-timeout: 1800<br>
><br>
> features.quota: off<br>
><br>
><br>
> Thanks in advanced.<br>
><br>
><br>
</div></div>> _______________________________________________<br>
> Gluster-users mailing list<br>
> <a href="mailto:Gluster-users@gluster.org">Gluster-users@gluster.org</a><br>
> <a href="http://supercolony.gluster.org/mailman/listinfo/gluster-users" target="_blank">http://supercolony.gluster.org/mailman/listinfo/gluster-users</a><br>
</blockquote></div><br></div>