<div dir="ltr"><div><div>>I was asking about reading data in same disperse set like 8+2 disperse
config if one disk is replaced and when heal is in process and when
client reads data which is available in rest of the 9 disks.<br><br></div>My use case is write heavy, we barely read data, so I do not know if read speed degrades during heal. But I know write speed do not change during heal.<br><br></div>How big is your files? How many files on average in each directory?<br></div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Apr 18, 2017 at 11:36 AM, Amudhan P <span dir="ltr"><<a href="mailto:amudhan83@gmail.com" target="_blank">amudhan83@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 dir="ltr"><div><br></div>I actually used this (find /mnt/gluster -d -exec getfattr -h -n trusted.ec.heal {} \; > /dev/null<br>) command on a specific folder to trigger heal but it was also not showing any difference in speed.<div><br></div><div>I was asking about reading data in same disperse set like 8+2 disperse config if one disk is replaced and when heal is in process and when client reads data which is available in rest of the 9 disks. </div><div><br></div><div>I am sure there was no bottleneck on network/disk IO in my case. </div><div><br></div><div><table style="border-collapse:collapse;width:583pt" border="0" cellspacing="0" cellpadding="0" width="777">
<colgroup><col style="width:583pt" width="777"> </colgroup><tbody><tr style="height:15pt" height="20">
<td style="height:15pt;width:583pt" width="777" height="20">I have tested 3.10.1 heal with disperse.shd-max-threads = 4. heal completed data size of 27GB in 13M15s. so it works well in a test environment but production environment it differs.<br><br></td>
</tr>
<tr style="height:15pt" height="20">
<td style="height:15pt" height="20"><br></td>
</tr></tbody></table><div><div class="h5"><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Apr 18, 2017 at 12:47 PM, Serkan Çoban <span dir="ltr"><<a href="mailto:cobanserkan@gmail.com" target="_blank">cobanserkan@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">You can increase heal speed by running below command from a client:<br>
find /mnt/gluster -d -exec getfattr -h -n trusted.ec.heal {} \; > /dev/null<br>
<br>
You can write a script with different folders to make it parallel.<br>
<br>
In my case I see 6TB data was healed within 7-8 days with above command running.<br>
<span class="m_7902750467696388854gmail-">>did you face any issue in reading data from rest of the good bricks in the set. like slow read < KB/s.<br>
</span>No, nodes generally have balanced network/disk IO during heal..<br>
<br>
You should make a detailed tests with non-prod cluster and try to find<br>
optimum heal configuration for your use case..<br>
Our new servers are on the way, in a couple of months I also will do<br>
detailed tests with 3.10.x and parallel disperse heal, will post the<br>
results here...<br>
<div class="m_7902750467696388854gmail-HOEnZb"><div class="m_7902750467696388854gmail-h5"><br>
<br>
On Tue, Apr 18, 2017 at 9:51 AM, Amudhan P <<a href="mailto:amudhan83@gmail.com" target="_blank">amudhan83@gmail.com</a>> wrote:<br>
> Serkan,<br>
><br>
> I have initially changed shd-max-thread 1 to 2 saw a little difference and<br>
> changing it to 4 & 8. doesn't make any difference.<br>
> disk write speed was about <1MB and data passed in thru network for healing<br>
> node from other node were 4MB combined.<br>
><br>
> Also, I tried ls -l from mount point to the folders and files which need to<br>
> be healed but have not seen any difference in performance.<br>
><br>
> But after 3 days of heal process running disk write speed was increased to 9<br>
> - 11MB and data passed thru network for healing node from other node were<br>
> 40MB combined.<br>
><br>
> Still 14GB of data to be healed when comparing to other disks in set.<br>
><br>
> I saw in another thread you also had the issue with heal speed, did you face<br>
> any issue in reading data from rest of the good bricks in the set. like slow<br>
> read < KB/s.<br>
><br>
> On Mon, Apr 17, 2017 at 2:05 PM, Serkan Çoban <<a href="mailto:cobanserkan@gmail.com" target="_blank">cobanserkan@gmail.com</a>> wrote:<br>
>><br>
>> Normally I see 8-10MB/sec/brick heal speed with gluster 3.7.11.<br>
>> I tested parallel heal for disperse with version 3.9.0 and see that it<br>
>> increase the heal speed to 20-40MB/sec<br>
>> I tested with shd-max-threads 2,4,8 and saw that best performance<br>
>> achieved with 2 or 4 threads.<br>
>> you can try to start with 2 and test with 4 and 8 and compare the results?<br>
><br>
><br>
</div></div></blockquote></div><br></div></div></div></div></div>
</blockquote></div><br></div>