<div dir="auto">Wonderful, it doesn't sound too bad then, I will be using three nodes with Kubernetes. Thanks! </div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, 9 Aug 2019, 13.45 Roland Kammerer, <<a href="mailto:roland.kammerer@linbit.com">roland.kammerer@linbit.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Thu, Aug 08, 2019 at 10:41:37PM +0300, Vito Botta wrote:<br>
> Hi,<br>
> <br>
> at the moment I am testing with 3 nodes that have 4 cores and 16 GB of<br>
> ram, and all works great. However since DRBD/Linstor have very little<br>
> overhead compared to other storage products I've tried with<br>
> Kubernetes, I was wondering whether 2 cores and 8 GB of ram would be<br>
> sufficient? <br>
<br>
LINSTOR/DRBD are not that resource hungry. Java is a but, but Rene got<br>
some pretty nice results with some minimal configurations. Rene?<br>
<br>
> Also, if I have three servers with volumes being replicated, how can I<br>
> avoid downtime when upgrading whenever a new release is out?<br>
<br>
That depends a bit what you want to update:<br>
<br>
The really nice thing about DRBD and LINSTOR are that the control plane<br>
is separated from the data plane. Which means you shut down LINSTOR<br>
completely and your data is still up, because DRBD continues to<br>
replicate data. So updating LINSTOR is easy, just do it :)<br>
<br>
If you need to update DRBD (i.e., the kernel modules), then you<br>
basically do this:<br>
- move all resources away from the node you want to update (i.e, no DRBD<br>
resource is Primary anymore)<br>
- drbdadm down all<br>
- rmmod<br>
- upgrade<br>
<br>
And this then you do the same for the next node. We are very careful<br>
that you can have mixed DRBD version until you finally upgraded your<br>
last storage node. Obviously you need to have at least 2 storage nodes,<br>
if you only have one diskful and one diskless node, then yeah, you will<br>
have downtime. And maybe you want to first provide "over redundancy" via<br>
an additional diskful replica, or you are fine that your N times<br>
redundant resource is only N-1 during you upgrade your storage. But that<br>
all depends on your use case.<br>
<br>
The important thing is that data and control plane.<br>
<br>
Best, rck<br>
_______________________________________________<br>
Star us on GITHUB: <a href="https://github.com/LINBIT" rel="noreferrer noreferrer" target="_blank">https://github.com/LINBIT</a><br>
drbd-user mailing list<br>
<a href="mailto:drbd-user@lists.linbit.com" target="_blank" rel="noreferrer">drbd-user@lists.linbit.com</a><br>
<a href="http://lists.linbit.com/mailman/listinfo/drbd-user" rel="noreferrer noreferrer" target="_blank">http://lists.linbit.com/mailman/listinfo/drbd-user</a><br>
</blockquote></div>