Shutting down a node

You can shut down a node if it becomes unresponsive or if support personnel direct you to do so as part of troubleshooting efforts.

Steps

  1. If the cluster contains four or more nodes, verify that the node to be shut down does not hold epsilon:
    1. Set the privilege level to advanced: set -privilege advanced
    2. Determine which node holds epsilon: cluster show
      Example

      The following example shows that "node1" holds epsilon:
      cluster1::*> cluster show
      Node                 Health  Eligibility   Epsilon
      -------------------- ------- ------------  ------------
      node1                true    true          true
      node2                true    true          false
      node3                true    true          false
      node4                true    true          false
      4 entries were displayed.
      
    3. If the node to be shut down holds epsilon, then remove epsilon from the node: cluster modify -node node_name -epsilon false
    4. Assign epsilon to a different node that will remain up: cluster modify -node node_name -epsilon true
    5. Return to the admin privilege level: set -privilege admin
  2. Use the system node halt command to shut down the node.
    If you do not specify the -skip-lif-migration parameter, the command attempts to migrate data and cluster management LIFs synchronously to another node prior to the shutdown. If the LIF migration fails or times out, the shutdown process is aborted, and ONTAP displays an error to indicate the LIF migration failure.

    You can manually trigger a core dump with the shutdown by using both the -dump parameter.

    Example
    The following example shuts down the node named "node1" for hardware maintenance:
    cluster1::> system node halt -node node1 -reason 'hardware maintenance'