2.2.2 Hot standby operation in the event of a host failure

(1) Resetting hosts

If a host failure occurs when host reset is to be performed, the HA Monitor in the standby system issues a request to reset the active system. It issues this request to the failure management processor on the failed host via the reset path. The failure management processor that manages the failed host then performs reset processing, such as termination of input/output operations. When the reset processing is completed, the failed host sends a notification to that effect to the standby system. The standby system then starts hot standby processing.

The following figure provides an overview of HA Monitor's hot standby operation in the event of a host failure.

Figure 2‒8: Overview of the hot standby operation in the event of a host failure (when host reset is used)

To Page Top

(2) Using SCSI reservation for shared disk

If a host failure occurs when SCSI reservation for shared disk is used, a reservation is forcibly obtained by the target host, and then the active server is started so that the shared disk is protected from data corruption that might result from concurrent write operations from multiple hosts.

If the hot standby operation occurs due to a monitoring path failure and the failed host detects a loss of reservation, planned termination is performed on the server.

The following figure provides an overview of HA Monitor's hot standby operation in the event of a host failure.

Figure 2‒9: Overview of the hot standby operation in the event of a host failure (when SCSI reservation for shared disk is used)

The following figure provides an overview of hot standby in the event of a monitoring path failure when SCSI reservation for shared disk is used.

Figure 2‒10: Hot standby in the event of a host failure (monitoring path failure when SCSI reservation for shared disk is used)

To Page Top

(3) Using the function for controlling hot standby based on the availability of LAN communications

If a host failure occurs while the function for controlling hot standby based on the availability of LAN communications is being used, whether LAN communication is available is checked. If the standby system confirms that LAN communication is available, the hot standby operation is performed. In the event of a monitoring path failure, HA Monitor ensures that a server in the system in which LAN communication is available will perform job processing.

The following figure provides an overview of HA Monitor's hot standby operation in the event of a host failure.

Figure 2‒11: Overview of the hot standby operation in the event of a host failure (when the function for controlling hot standby based on the availability of LAN communications is used)

The following figure provides an overview of the hot standby operation in the event of a monitoring path failure when the function for controlling hot standby based on the availability of LAN communications is used.

Figure 2‒12: Hot standby operation in the event of a host failure (monitoring path failure when the function for controlling hot standby based on the availability of LAN communications is used)

To Page Top