[ClusterLabs] Adding node in existing cluster pcs constraint not setting properly

S Sathish S s.s.sathish at ericsson.com
Thu May 14 04:41:12 EDT 2020

Hi Team,

Recently during installation at the customer environments we came  across an issue where o/p of "pcs status" of the node to be added in the cluster was as follows:

Command triggered at NodeC

# sudo pcs status
Cluster name: XXX

No stonith devices and stonith-enabled is not false

Stack: corosync
Current DC: NODE C (version 2.0.2-744a30d655) - partition WITHOUT quorum
Last updated: Wed May 13 14:49:40 2020
Last change: Wed May 13 14:39:04 2020 by hacluster via crmd on NODE C

3 nodes configured
0 resources configured

Node A: UNCLEAN (offline)
Node B: UNCLEAN (offline)
Online: [ NODE C ]

No resources

  Daemon Status:
  corosync: active/enabled
  pacemaker: active/enabled
  pcsd: active/enabled

Problem Statement: Node A and B were the nodes that have already formed a cluster and Node C was the node that had to be added as a new member into the cluster. The above output was fetched from NodeC.

The authorization and addition of NodeC in our case is done from NodeB.

On TroubleShooting we found that the necessary pcs constraints stonith setting did not get set for NodeC and hence the above reported warning was seen.

In our project from the start, we configure the resources and set the necessary pcs constraints stonith setting of the new node(NodeC) will be triggered from existing cluster (NodeB).

Once the authorization is complete is it the right procedure to set the constraints of NodeC in NodeB (or)  the constraints need to be set at NodeC itself.

Kindly let us know if we have followed the right procedure. If not, kindly suggest an alternative.

Thanks and Regards,
S Sathish S
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.clusterlabs.org/pipermail/users/attachments/20200514/42d774f6/attachment.htm>

More information about the Users mailing list