Verifying cluster failover

This procedure simulates a failover.
  1. Log in to the primary node as root, or as a user with superuser privileges.
  2. Enable the DRBD secondary node.
    1. Take the secondary node out of standby mode.

      Replace Secondary-Public-Name with the public hostname of the secondary node:

      pcs cluster unstandby Secondary-Public-Name
    2. Monitor the status of the secondary node.
      pcs status
      Do not continue until the status of the secondary node is Online.
  3. Verify that DRBD has completely synchonized all three volumes on the secondary node.
    drbd-overview
    Example result:
    0:serviced-dfs/0  Connected Primary/Secondary UpToDate/UpToDate
    1:serviced-dfs/1  Connected Primary/Secondary UpToDate/UpToDate
    2:serviced-dfs/2  Connected Primary/Secondary UpToDate/UpToDate
  4. Force a failover.
    Pacemaker initiates a failover when the primary node is put in standby mode.

    Replace Primary-Public-Name with the public hostname of the primary node:

    pcs cluster standby Primary-Public-Name
  5. Monitor the cluster status.
    pcs status
    Repeat the preceding command until all resources report a status of Started. Resolve any issues before continuing.
  6. Restore the cluster.

    Replace Primary-Public-Name with the public hostname of the primary node:

    pcs cluster unstandby Primary-Public-Name