Generic ARK for HULFT Overview

HULFT resources have four operations.

Bringing a Resource In Service (restore)

To set the HULFT resource to ISP (In Service Protected), the HULFT service is started using the HULFT executable binary program. This operation corresponds to when the startup process for the HULFT resource is executed via the LifeKeeper GUI client or CUI. If the startup process fails, the HULFT resource enters OSF (Out-of-service, failed). The same process is performed during switchover and failover operations. For details on this operation, refer to the Generic ARK for HULFT Script Specification.

Taking a Resource Out of Service (remove)

To place the HULFT resource into OSU (Out-of-service, unimpaired) state, the HULFT service is stopped using the HULFT executable binary program. This operation corresponds to when the stop process for the HULFT resource is executed via the LifeKeeper GUI client or CUI. If the stop process fails, the behavior differs depending on the LKHULFT_FORCE_REMOVE parameter. By default, the HULFT resource enters OSU state. For details on the parameters, refer to the Generic ARK for HULFT Parameter List described. The same process is performed during switchover and failover operations. For details on this behavior, refer to the Generic ARK for HULFT Script Specification.

Monitoring (quickCheck)

When the HULFT resource is in ISP mode, it verifies that the HULFT service is operating normally. If the monitoring process does not confirm correct operation, the state transitions to the recovery process. For details on this behavior, refer to the Generic ARK for HULFT Script Specification.

Recovery (recover)

This operation is executed only when a failure is detected during monitoring. On the node where the failure was detected during monitoring, the shutdown process (remove) is performed, followed by the startup process (restore). If the recovery process fails, a failover to the standby system is initiated. For details on this operation, refer to the Generic ARK for HULFT Script Specification. Details of the failover operation are described below.

Details of Failover Operation

The script is designed to switch in cases such as the following. HULFT resources are assumed to have restore scripts, remove scripts, quickCheck scripts, and recover scripts specified for each.

• HULFT Resource Failure
  1. When a failure is detected in a HULFT resource, it will behave as follows.
  2. Detects failures using the quickCheck process in HULFT resources.
     The HULFT resource recovery process is executed to perform recovery operations.
     • If the HULFT resource recovery process succeeds, the service will return to its normal state.
     • If the HULFT resource recovery process fails, it will fail over to the passive node.

• Node Failure
  If all communication paths are in a DEAD state and the Safety Check fails, failover will begin on the passive node. During this failover, the passive node sequentially performs startup processing for each resource. The behavior during HULFT resource startup at this time is to perform the startup processing. For details, refer to “Bringing a Resource In Service (restore)”.