7.9.1 Forcibly performing hot-standby switchover
The OS operation in the active system might become unstable if the CPU load, memory load, and disk I/O load become high. In such a case, the active server might be unable to execute job processing. Note that if HA Monitor operates normally even when the OS operation is unstable, HA Monitor might not perform hot-standby switchover without detecting a host failure. In such a case, you can perform hot-standby switchover by forcibly terminating the OS on which the active server is running. This operation is called forced hot-standby switchover. You can use forced hot-standby switchover in the following cases:
-
When you cannot log in to the OS of the active system
If the OS operation of the active system is unstable and you cannot log in to the active system, you might be unable to perform planned hot-standby switchover from the active system. In such a case, you can perform hot-standby switchover by executing the monswap command with the -f or -f -g option in the standby system.
-
When attempted hot-standby switchover does not finish
Planned hot-standby switchover attempted from the active system or hot-standby switchover attempted when a failure is detected by the server monitoring command might not be completed for some reason. For example, this situation occurs if it takes time to terminate the server or to disconnect shared resources in the active system. If disconnection of shared resources or another process takes time, the KAMN310-I message (hot-standby switchover start message) is not output in the standby system. In such a case, you can perform hot-standby switchover by executing the monswap command with the -f or -f -g option in the standby system.