<div dir="ltr">
<p class="">Hi,</p>
<p class=""><br></p>
<p class="">We have a set of 4 gluster nodes, all in replicated (design?)</p>
<p class="">We use it to store our qcow2 images for kvm. These images have a variable IO, though most of them are for reading only.</p>
<p class=""><br></p>
<p class="">I tried to find some documentation re. performance optimization, but it's either unclear to me, or I couldn't find much.. so I'd copied from the internet and tried to adjust the config to our needs, but I'm sure it's not optimized.</p>
<p class=""><br></p>
<p class="">We're using 3.3.1 on top of XFS.</p>
<p class="">The qcow images are about 30GB (a couple of 100GB).</p>
<p class=""><br></p>
<p class="">Can someone please tell what would be the best paramenter for performance to look at?</p>
<p class=""><br></p>
<p class="">here is volume info:</p>
<p class=""><br></p>
<p class="">Volume Name: images</p>
<p class="">Type: Replicate</p>
<p class="">Volume ID: </p>
<p class="">Status: Started</p>
<p class="">Number of Bricks: 1 x 4 = 4</p>
<p class="">Transport-type: tcp</p>
<p class="">Bricks:</p>
<p class="">Brick1: vmhost2:/exports/1</p>
<p class="">Brick2: vmhost3:/exports/1</p>
<p class="">Brick3: vmhost5:/exports/1</p>
<p class="">Brick4: vmhost6:/exports/1</p>
<p class="">Options Reconfigured:</p>
<p class="">performance.cache-max-file-size: 1GB</p>
<p class="">nfs.disable: on</p>
<p class="">performance.cache-size: 4GB</p>
<p class="">performance.cache-refresh-timeout: 1</p>
<p class="">performance.write-behind-window-size: 2MB</p>
<p class="">performance.read-ahead: on</p>
<p class="">performance.write-behind: on</p>
<p class="">performance.io-cache: on</p>
<p class="">performance.stat-prefetch: on</p>
<p class="">performance.quick-read: on</p>
<p class="">performance.io-thread-count: 64</p>
<p class="">performance.flush-behind: on</p>
<p class="">features.quota-timeout: 1800</p>
<p class="">features.quota: off</p>
<p class=""><br></p>
<p class="">Thanks in advanced.</p></div>