<div dir="ltr">I'm on the fence about the amount of time it will take to degrade and rebuild a RAID6 at 16 drives (x2 systems). <div><br></div><div>Anyone against the idea of:</div><div>Backup data friday night through saturday morning</div>
<div style>stop drbd and heartbeat on node2</div><div style>replace all drives on node2</div><div style>build raid 6 and match setup/sizes from node1</div><div style>initialize metadata, etc.</div><div style>start drbd and heartbeat</div>
<div style>let it sync</div><div style>make node2 primary</div><div style>repeat steps for node1</div><div style><br></div><div style>How does this sound? You'll be risking the data only during syncing process but you would be doing that for 12-14 drives (per system) to tell them to rebuild the RAID. Either way you have tape or some form of backup...DRBD is not a backup :) There I said it.</div>
<div style><br></div><div style>Let's start off with, you're risking your data during this process no matter what way you go about it. I had Seagate ES drives die during a rebuild of node2 because the syncing 100% reading from node1 and 100% writing to node2 was just too much for the older drives to handle. I think the mixing of drives works in theory but it something goes awry in the process you'll do your best to get it running and most likely stop the upgrade process temporarily. This will leave you with some 1TB and some 2TB disks in a running state with fear of further breaking it,etc.<br>
</div><div style><br></div><div style><br></div></div><div class="gmail_extra"><br><br><div class="gmail_quote">On Thu, Jan 31, 2013 at 11:20 AM, Adam Goryachev <span dir="ltr"><<a href="mailto:mailinglists@websitemanagers.com.au" target="_blank">mailinglists@websitemanagers.com.au</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div text="#000000" bgcolor="#FFFFFF"><div><div class="h5">
<div>On 01/02/13 02:58, Marcelo Pereira
wrote:<br>
</div>
<blockquote type="cite">
<div>Hello Everyone,</div>
<div><br>
</div>
<div>I'm about to perform an upgrade on my servers and I was
wondering how to do that.</div>
<div><br>
</div>
<div>Here is the scenario:</div>
<div><br>
</div>
<div>Server A has 16x 1Tb hard drives, under RAID-6.</div>
<div>Server B has 16x 1Tb hard drives, under RAID-6.</div>
<div><br>
</div>
<div>And both are in sync, using DRBD.</div>
<div><br>
</div>
<div>I though about replacing the hard drives for 2Tb units, one
by one.</div>
<div><br>
</div>
<div>So, on each run, I would:</div>
<ul>
<li>Remove a 1Tb disk</li>
<li>Add a 2Tb disk</li>
<li>Wait for it to rebuild the RAID</li>
</ul>
<div>After replacing ALL disks, I would expand the RAID unit, on
each server.</div>
<div><br>
</div>
<div>However, I was wondering how DRBD would "like" this
procedure.</div>
<div><br>
</div>
<div>I know that, before "expanding" the RAID, the cluster size,
and the block numbers would remain the same, as I would be
"wasting" the extra space on the newly added drives.</div>
<div><br>
</div>
<div>So, after "both" servers have all the drives replaces, and
the RAID is properly rebuild. Would that be a problem to expand
it? How would DRBD handle it?</div>
<div><br>
</div>
<div>I will appreciate any comment or suggestion here.</div>
</blockquote></div></div>
DRBD will work perfectly...<br>
<br>
You probably need to do the following:<br>
1) Pull one drive and replace (you could do one on each server at
the same time, although better/safer to do one server at a time)<br>
2) Wait for rebuild to complete<br>
3) Repeat for all disks on BOTH servers<br>
4) Resize the RAID array on each server<br>
5) Resize DRBD (see the fantastic online manual for your version of
DRBD for the details)<br>
6) Resize the underlying filesystem or whatever<br>
<br>
BTW, depending on your kernel version, and/or RAID (I'm assuming
linux software raid), you might like to query the linux-raid list to
see if you can ADD the new drive, tell md that this new drive is
replacing drive X, this way you avoid degrading the RAID array,
hence lose less performance during the rebuild, and have a lower
risk of disk failure and especially URE (Unrecoverable Read Error)
during the rebuilds.<br>
<br>
Regards,<br>
Adam<span class="HOEnZb"><font color="#888888"><br>
<br>
<br>
<pre cols="72">--
Adam Goryachev
Website Managers
<a href="http://www.websitemanagers.com.au" target="_blank">www.websitemanagers.com.au</a>
</pre>
</font></span></div>
<br>_______________________________________________<br>
drbd-user mailing list<br>
<a href="mailto:drbd-user@lists.linbit.com">drbd-user@lists.linbit.com</a><br>
<a href="http://lists.linbit.com/mailman/listinfo/drbd-user" target="_blank">http://lists.linbit.com/mailman/listinfo/drbd-user</a><br>
<br></blockquote></div><br></div>