In a disaster recovery environment, upgrade gets stuck during node evacuation stage as VVRInfra_Grp cannot be brought down
During upgrade, all the service groups are brought offline on each node and then the node is evacuated from the cluster. The upgrade process waits for all service groups to be offline for the evacuation to succeed on that node. The VVRInfra_Grp service group may be stuck and can fail to go offline because of the following reasons:
A resource of type RVGShared is present, which tries to restart vradmind service as soon as it finds the that service is not running.
/opt/VRTSnas/scripts/rep/nso_replication.sh auto_resyncdis in running state and restarts the vradmind service.
(IA-40311)
Workaround:
Login to Appliance shell using admin user credentials on any node in the cluster except for the node on which the evacuation is not hung.
Enter the support elevate command to elevate to maintenance mode.
Enter the maintenance user account password to log into the maintenance mode.
On the bash shell, execute the following commands in the specified order.
# /opt/VRTSvcs/bin/haconf -makerw
# /opt/VRTSvcs/bin/hares -modify VVRInfra_Grp-vvr_daemon StartProgram "/usr/sbin/vxstart_vvr start && /bin/touch /var/run/VRTSvcs/vvr_service"
# /opt/VRTSvcs/bin/hares -modify VVRInfra_Grp-vvr_daemon StopProgram "/usr/sbin/vxstart_vvr stop && /bin/rm -f /var/run/VRTSvcs/vvr_service"
# /opt/VRTSvcs/bin/hares -modify VVRInfra_Grp-vvr_daemon CleanProgram "/usr/sbin/vxstart_vvr stop && /bin/rm -f /var/run/VRTSvcs/vvr_service"
# /opt/VRTSvcs/bin/haconf -dump -makero
Log into the maintenance mode and execute the following command on the node on which the evacuation is hung.
# rm -f /var/run/VRTSvcs/vvr_service