Problem Symptom
A management node is restarted during an upgrade and the upgrade is paused. After processes are restored, DeviceManager fails to be logged in.
Problem Diagnosis
Confirm that a node has been restarted during an upgrade.
Logging in to DeviceManager fails.
Check the /var/log/deploy/clouda/script/action.log log file on the active management node. The last component that was being upgraded is DeviceManager and the upgrade was not complete.
notice：
Atomic tasks during the DeviceManager component upgrade:
offline > stop > upgrade > start > check status > online
Causes
The upgrade is not complete due to a node restart. After the CloudA process is started, it detects that an atomic task fails and the upgrade process is suspended. The active/standby management node switchover is disabled during the upgrade, and the upgrade process is not complete on DeviceManager. Therefore, you may fail to log in to DeviceManager.
Solution
Log in to the active management node.
Query the last atomic task that is suspended during the upgrade in the /var/log/deploy/clouda/script/action.log log file.
Go to the /opt/fusionstorage/repository/deploymanager_pkg/micro directory. Find the devicemanager folder of the target version, and go to the action folder.
Execute the remaining atomic tasks in sequence from the interrupted atomic tasks. For example, the log shows that the node is restarted during the upgrade atomic task. After the node is restored, run the following commands in sequence:
Command 1: bash appctl.sh upgrade
Command 2: bash appctl.sh start
Command 3: bash appctl.sh checkstatus
Command 4: bash appctl.sh online
Log in to DeviceManager. If the login succeeds, continue the upgrade on DeviceManager.
Check After Recovery
N/A
Suggestion and Summary
N/A
Applicable Versions
OceanStor 100D 8.0.3; OceanStor Pacific series 8.1.0; OceanStor Atlantic series 8.1.0
