CONTROLLER TAKEOVER COMPLETE AUTOMATIC - Communication Error ALERT in new cluster configuration
Applies to
- AFF A20
- Initial cluster configuration / setup
Issue
- Unexpected takeover event reported. Example:
HA Group Notification from node_name (CONTROLLER TAKEOVER COMPLETE AUTOMATIC - Communication Error) ALERT
- From the partner node, the following alerts are mentioned:
[node_name: statd: cf.takeover.disabled:alert]: HA mode, but takeover of partner is disabled due to reason : unsynchronized log.
with:
[node_name: ThreadHandlerun: cf.fsm.clam.reqPartnerShtdwn:alert]: CLAM requests graceful shutdown of the HA partner to initiate a takeover while NVLOG is out of sync. Cluster and HA connectivity is down.
...
[node_name: cf_main: cf.fsm.takeover.on.reboot:info]: Failover monitor: One node initiated automatic takeover after detecting that its partner node is rebooting.
...
[node_name: shutdown_thread0: ha.localNodeShutDown:notice]: Shutdown of the local node has been initiated with inhibit_takeover set to FALSE.
- The node down is not able to boot into "
Waiting for giveback...
" status. - An unexpected unsupported cluster network switch (Switch documentation for ONTAP hardware systems) is detected, under the "
::> system switch ethernet show
" ONTAP CLI output, or the CSHM-SWITCH-CONFIG.XML AutoSupport section. Example:
Device Name switch_name (aa:bb:cc:dd:ee:ff)
IP Address 192.168.0.1
Model to display OTHER
Switch Network cluster-network
Software Version switch_name_firmware...
Reference Config File Version NA
SNMP Version SNMPv2c
...
Serial Number of the Device Unknown
...
- The same device is detected as connected to the Cluster/HA physical ports for this platform: e4a and e4b.