<div dir="ltr"><div><div><div><div>Hi, <br><br></div>I've been using drbd for a few months now and I am far from an expert. So please excuse me for any stupid question.<br><br></div>I have the following task: <br>- Cluster of 3 nodes(Node1 and Node2 - main nodes, Node3 - backup node). I know I should use Stacked devices. <br>
- Managed by Pacemaker(version 1.1.8). <br>- Drbd(version 8.3) replication between the 3 nodes. <br>- Node3 is backup. It should stay slave until both Node1 and Node2 are down. In this case, Node3 should be able to bring up all the resources automatically. Once Node1 and/or Node2 are up, Node3 should pass the resources back to them.</div>
<br></div><div><div>I searched the web, did lots of setups and tests with the existing drbd resource agent. No success. It seems that the default drbd RA is unable to handle this configuration.<br></div><div>So I decided to extend it for my particular case. I did this and succeeded, but the setup seems heavy, complex and difficult to track. And sometimes crashes.<br>
</div><div><br>I don't mind discussing my setup. <br>But for now I'd like to ask this:<br></div><div>- Have you approached such configuration
before? Did you come up with solution?<br></div><div>- Is there working configuration for such setup? I mean using
the existing drbd resource agent<br></div><div>- Can I ask for an example? Guidelines?<br> </div><div><br></div><div>Thanks in advance.<br></div><div>Greetings, Nedelin.<br></div></div></div>