1.4 Hot-standby switchover methods
The following hot-standby switchover methods can be used with HA Monitor:
-
Hot-standby switchover based on hybrid fencing
-
Hot-standby switchover based on a host reset
-
Hot-standby switchover based on SCSI reservation for shared disk
-
Hot-standby switchover based on the availability of LAN communications
For HA Monitor, we recommend hot-standby switchover based on hybrid fencing. This method has the following advantages:
-
Even if host reset cannot be performed due to, for example, a failure of the failure management processor (SVP or VMware ESXi), the system can continue to perform hot-standby switchover by using a SCSI-reserved shared disk. Because the reliability of automatic hot-standby switchover in the event of a host failure is higher than other methods, system availability can be improved.
-
The failed host can be stopped and isolated completely by resetting the host, which prevents the same server from starting on multiple hosts as active servers. Thus, safe hot-standby switchover can be performed.
-
A dump can be obtained if an OS slowdown occurs.
If hot-standby switchover based on hybrid fencing cannot be selected, hot-standby switchover based on host reset is recommended. Hot-standby switchover based on host reset has the following advantages:
-
The failed host can be stopped and isolated completely by resetting the host, which prevents the same server from starting on multiple hosts as active servers. Therefore, safe hot-standby switchover can be performed.
-
A dump can be obtained if OS slowdown occurs.
Note, however, that the hot-standby switchover methods that can be selected differ depending on the system configuration and requirements. Select a hot-standby switchover method based on the information in the following table.
System configuration (server model) |
Hot standby method |
---|---|
BladeSymphony or HA8000 |
|
Non-Hitachi PC/AT-compatible system |
|