5.2.2 Considerations for designing a configuration
When you design a hot-standby configuration, you must ensure the reliability of the monitoring path and reset path.
The reliability of the monitoring and reset paths is important for detecting host failures. If all of the monitoring and reset paths fail at the same time, HA Monitor will not be able to detect the failure on the reset path using a health check, and instances of the same server might end up running on both the active system and the standby system. Therefore, to ensure the reliability of the monitoring path and reset path, do the following:
-
Always use multiple monitoring paths, and dedicate at least one of these to monitoring only.
-
Always perform health checks on the monitoring and reset paths.
-
In the event of failure on a monitoring or reset path, promptly eliminate the cause of the failure.