Problem Symptom
Before the first DeployManager upgrade, node A is the standby FSM node and node B is the active FSM node.
Before the first DeployManager upgrade, an active/standby switchover is triggered and node A becomes the active FSM node. However, the CloudA upgrade fails. During environment recovery, an active/standby switchover is triggered again unexpectedly and node A becomes the standby FSM node. As a result, an error is reported during the second DeployManager upgrade on node A, indicating that the current node is not the active FSM node.
[2020-06-28 11:21] is_active [Line:146][Error] current node ha role is not active.You can upgrade active node only[2020-06-28 11:21] current node ha role is not active,upgrade cloude error[2020-06-28 11:21] DeployManager upgrade result is :1[2020-06-28 11:21] [ERROR]DeployManager upgrade failed![2020-06-28 11:21] The upgrade is failed
Problem Diagnosis
Log in to node A and check the package.log log file. The log file shows that node A is not the active FSM node. On DeviceManager, two times of active/standby switchover are recorded. However, in normal cases, only one switchover will be recorded.
Causes
After the environment is recovered upon a CloudA upgrade failure, an abnormal active/standby switchover is triggered and node A becomes the standby FSM node. Therefore, the second DeployManager upgrade fails because CloudA must be upgraded on the active FSM node.
Solution
Ensure that the DeployManager upgrade environment is normal.
Log in to node B and run the following command to perform an active/standby switchover:
cd /opt/dfv/oam/oam-u/ha/ha/module/hacom/tools && ./ha_client_tool --cancelforbidswitch --name=upgrade && ./ha_client_tool --switchover --name=upgrade
Check the alarms on DeviceManager. After an active/standby switchover record is generated, try the upgrade again on node A.
