DRBDManage with Pacemaker on CentOS 7

Question

I have 2 CentOS 7.6 nodes which I have configured with drbd, drbdmanage and pacemaker for active/passive failover.

during a failure my vip and drbd service failover from primary to secondary the next node but I couldn't connect to drbdmanage to inspect/manage the cluster.

Since pacemaker has resource type for drbdmanage I assumed that I should be able to manage and failover the service to the surviving node.

I would like to know whether it's possible to failover drbdmanage after a node failure?

score 5 · Accepted Answer · answered Sep 12 '19 at 07:29

5

DRBDmanage will not work correctly in “pure” 2-node setup, because of lost quorum. drbdmanage will work properly in following cases:

1 node
2 nodes and both are online. Otherwise the quorum will not exist.
2 nodes + witness node https://docs.linbit.com/docs/users-guide-9.0/#s-configuring-quorum-tiebreaker
2+ nodes and with node majority

If you need “pure” 2 node high available storage take a look at solutions, which have been designed for it (like Starwind vsan free)

answered Sep 12 '19 at 07:29

batistuta09

8,981
10
23

1

Thank you for the confirmation. I will look at the option with configuring it with witness node as a tiebreaker. – Sajirupee Sep 12 '19 at 14:02
1

Just a side note, DRBDManage is deprecated by LINBIT and should no longer be used. Check out LINSTOR if you need to dynamically provision DRBD devices. – Matt Kereczman Sep 12 '19 at 17:08

DRBDManage with Pacemaker on CentOS 7

1 Answers1