Verifying cluster startup

This procedure verifies the initial configuration by attempting to start the resources on one node only. With the other node in standby mode, Pacemaker does not automatically fail over to the other node.
  1. Log in to the primary node as root, or as a user with superuser privileges.
  2. In a separate window, log in to the secondary node as root, or as a user with superuser privileges.
  3. On the primary node, determine which node is the primary DRBD node.
    pcs status
    Example result:
    Cluster name: serviced-ha
    Last updated: Mon Feb 22 11:37:58 2016  Last change: Mon Feb 22 11:35:19 2016 by root via crm_attribute on Secondary-Public-Name
    Stack: corosync
    Current DC: Primary-Public-Name (version 1.1.13-a14efad) - partition with quorum
    2 nodes and 10 resources configured
    Node Primary-Public-Name: standby
    Node Secondary-Public-Name: standby
    Full list of resources:
     Master/Slave Set: DFSMaster [DFS]
     Stopped: [ Primary-Public-Name Secondary-Public-Name ]
     Resource Group: serviced-group
         serviced-isvcs  (ocf::heartbeat:Filesystem):  Stopped
         serviced-volumes  (ocf::heartbeat:Filesystem):  Stopped
         serviced-lvm  (ocf::heartbeat:LVM):  Stopped
         serviced-storage  (ocf::zenoss:serviced-storage):  Stopped
         VirtualIP  (ocf::heartbeat:IPaddr2):  Stopped
         docker  (systemd:docker):  Stopped
         nfs  (systemd:nfs):  Stopped
         serviced  (ocf::zenoss:serviced):  Stopped
    PCSD Status:
      Primary-Public-Name: Online
      Secondary-Public-Name: Online
    Daemon Status:
      corosync: active/disabled
      pacemaker: active/enabled
      pcsd: active/enabled

    The line that begins with Current DC identifies the primary node. Review all of the command output for errors.

  4. Start DRBD.
    1. On the secondary node, enter the following command:
      drbdadm up all
    2. On the primary node, enter the following commands:
      drbdadm up all && drbdadm primary serviced-dfs
  5. Start cluster resources.
    You can run pcs commands on either node.
    pcs cluster unstandby Primary-Public-Name
  6. Monitor the status of cluster resources.
    watch pcs status
    Monitor the status until all resources report Started. Resolve any issues before continuing.