A few nodes do not shutdown by halting whole cluster
Applies to
- ONTAP 9
- Over 4 nodes cluster
Issue
- When ONTAP performs the following command for maintenance shutdown, a few nodes do not shutdown.
Example:
::>system node halt -node * -inhibit-takeover true -skip-lif-migration-before-shutdown true
Warning: Are you sure you want to halt node "node1"? {y|n}: y
Waiting for a quorum membership change to complete.
Waiting for a quorum membership change to complete.
...
- The following events appear on EMS log.
Example:
Fri Jan 04 21:31:01 JST [node1: shutdown_thread0: kern.shutdown:notice]: System shut down because : "D-blade Halting".
Fri Jan 04 21:32:20 JST [node2: shutdown_thread0: kern.shutdown:notice]: System shut down because : "D-blade Halting".
Fri Jan 04 21:33:24 JST [node3: kltp: clam.node.ooq:EMERGENCY]: Node (name=node2, ID=1003) is out of "CLAM quorum" (reason=quorum update).
Fri Jan 04 21:33:24 JST [node4: kltp: clam.node.ooq:EMERGENCY]: Node (name=node2, ID=1003) is out of "CLAM quorum" (reason=quorum update).
Fri Jan 04 21:33:24 JST [node5: kltp: clam.node.ooq:EMERGENCY]: Node (name=node2, ID=1003) is out of "CLAM quorum" (reason=heartbeat failure).
Fri Jan 04 21:33:24 JST [node6: kltp: clam.node.ooq:EMERGENCY]: Node (name=node2, ID=1003) is out of "CLAM quorum" (reason=quorum update).
Fri Jan 04 21:33:24 JST [node7: kltp: clam.node.ooq:EMERGENCY]: Node (name=node2, ID=1003) is out of "CLAM quorum" (reason=quorum update).
Fri Jan 04 21:33:24 JST [node8: kltp: clam.node.ooq:EMERGENCY]: Node (name=node2, ID=1003) is out of "CLAM quorum" (reason=quorum update).