Steps to reproduce:
- Setup 2-node cluster with SBD
- on node2, disable iscsi initiator service
- Restart node2
- on node2, start iscsi initiator service
- Run
crm cluster start on node2
# crm cluster start
INFO: Starting pacemaker.service on node2
ERROR: The cluster stack failed to start on node2
corosync.service on node2 is running now, at node1, crm_mon shows node2 is pending
- Run
crm cluster start again on node2, success