2.3.3 Detecting server failures (in the monitor mode)
When the server is in the monitor mode, how server failures are detected depends on whether APIs are issued.
- Organization of this subsection
(1) When APIs are not issued
HA Monitor cannot detect server failures because it does not monitor servers. HA Monitor detects only host failures.
If a server failure occurs in the monitor mode, the user must take the following action:
-
If hot standby switchover to the standby system is necessary
Stop the active server and execute the server host hot-standby switchover command (monswap command) in the active system to perform planned hot-standby switchover.
-
If hot standby switchover to the standby system is not necessary
Restart the server in the active system
If you use the server monitoring command beforehand to create processing for monitoring for server failures, HA Monitor can perform the hot standby operation in the event of a server failure. For details about the server monitoring command, see 3.2.1 Monitoring a server in the monitor mode (for the active server).
(2) When APIs are issued
If the monitor-mode program management function is used to issue APIs from the UAP, HA Monitor monitors the UAP to detect server slowdown.
HA Monitor detects server failures caused by the UAP by monitoring the UAP's processes and operation reports. You can choose to monitor both the UAP's processes and operation reports or you can choose to monitor only the UAP's processes.
In the event of a UAP failure, HA Monitor can also restart the UAP by issuing a program restart command. The program restart command must have been created in advance by the user.
For details about monitoring UAPs, see 3.6.1 Monitoring UAPs.