InfoScale™ 9.0 Cluster Server Agent for Hitachi TrueCopy/HP-XP Continuous Access Configuration Guide - Windows

online

If the state of all local devices is read-write enabled, the agent makes the devices writable by creating a lock file on the local host.

If one or more devices are not in a writable state, the agent runs the horctakeover command to enable read-write access to the devices.

For S-VOL devices in any state other than SSWS/SSUS/SMPL, the agent runs the horctakeover command and makes the devices writable. The time required for failover depends on the following conditions:

The agent considers P-VOL devices writable and takes no action other than going online, regardless of their status.

If the S-VOL devices are in the COPY state, the agent runs the horctakeover command after one of the following:

If the S-VOL devices are in the PAIR state, the agent runs the pairdisplay command without the -l option to retrieve the state of the remote site devices. If it finds that the P-VOL devices are in PAIR state, the agent proceeds with the failover. But if the remote RAID manager is down, then the agent honors the SplitTakeover attribute configuration before performing failover.

See About the Hitachi TrueCopy/Hewlett-Packard XP Continuous Access agent's online function.

If the RAID manager is not up for an instance, the agent runs the horcmstart command to bring the RAID manager online.

offline

The agent removes the lock file that was created for the resource by the online entry point. The agent does not run any TrueCopy commands because taking the resource offline is not indicative of an intention to give up the devices.

monitor

Verifies the existence of the lock file to determine the resource status. If the lock file exists, the agent reports the status of the resource as online. If the lock file does not exist, the agent reports the status of the resource as offline.

Based on other attribute values, the monitor entry point examines the state of the devices or the state of the replication link between the arrays.

open

Removes the lock file from the host on which this entry point is called. This functionality prevents potential concurrency violation if the group fails over to another node.

Note that the agent does not remove the lock file if the agent starts after the following command:

hastop<-all | -local> -force

clean

Determines whether if it is safe to fault the resource if the online entry point fails or times out. The main consideration is whether a management operation was in progress when the online thread timed out and it was killed. If a management operation was in progress, it could potentially leave the devices in an unusable state.

info

Reports the current role and status of the devices in the device group. This entry point can be used to verify the device state and to monitor dirty track trends.

action

action\PreSwitch

Ensures that the remote site cluster can come online during a planned failover within a GCO configuration without data loss. The VCS engine on the remote cluster invokes the PreSwitch action on all the resources of the remote Service Group during a planned failover using the hagrp -switch command. For this, the PreSwitch attribute must be set to 1. The option -nopre indicates that the VCS engine must switch the servicegroup regardless of the value of the PreSwitch service group attribute.

If running the PreSwitch action fails, the failover should not occur. This minimizes the application downtime and data loss .

For more information on the PreSwitch action and the PreSwitch feature in the VCS engine, refer to the Cluster Server Administrator's Guide.

action\vxdiske

Reports the mapping between the physical disk name and the volume manager disk name for all connected disks.