Returning production to h1 in a two Metro Mirror pair configuration

After a swap to h2 as a result of a site h1 failure, it might be determined that site h1 is operational and available.

About this task

An alternative to creating the h3 and h2 volume pairs is to suspend the h1 and h3 pairs and issue a failback command to the h2 to h1 volumes. To return to the original configuration, another failover and failback command is issued to the h2 to h1 volumes. However, this scenario does not use the Multi-Target PPRC Incremental Resync capabilities.

This scenario starts with the state that exists after the failover of h3 to h1, which is described in Failing of primary site, h1.

After a failure at the h1 site, production can be swapped to the h2 site using the procedures that are described in Two Metro Mirror pair configuration.

Figure 1 shows a configuration after a failure of the h1 site and a failover to h2.
Figure 1. Configuration after a failure of the h1 site and a failover to h2
The graphic a configuration after a failure at h1 and a failover to h2

Use the following procedure to resume replication to h1.

Procedure

  1. Failback from h2 to h1.
    1. Issue a failback establish command to the h2 to h1 volumes.
      This is a failback command without Multi-Target PPRC modifications.
      TSO/E command DS CLI command ANTRQST API command ICKDSF command
      CESTPAIR ACTION(FAILBACK) failbackpprc REQUEST=PESTPAIR ACTION=FAILBACK PPRCOPY ESTPAIR FAILBACK
      Figure 2. Configuration after a failback of the h2 to h1 volumes
      The graphic a configuration after a failure at h1 and a failover to h2
  2. Failover from h1 to h2
    1. Issue a failover command to the h1 to h2 volumes.
      TSO/E command DS CLI command ANTRQST API command ICKDSF command
      CESTPAIR ACTION(FAILOVER) failoverpprc REQUEST=PESTPAIR ACTION=FAILOVER PPRCOPY ESTPAIR FAILOVER

      After a failover from the h1 to h2 volumes, production I/O returns to site h1 as shown in Figure 3.

      Volume h1 is now a suspended primary to both h2 and h3 volumes. There is no active mirroring during the suspension.

      Figure 3. Configuration after a failover of the h1 to h2 volumes
      The graphic a configuration failover from h1 to h2
  3. Failback from h1 to h2
    1. Issue a failback command to the h1 to h2 volumes.
      This process resumes a Metro Mirror relationship for the volume pair and restarts transferring data from the h1 to h2 volumes.
      TSO/E command DS CLI command ANTRQST API command ICKDSF command
      CESTPAIR ACTION(FAILBACK) failbackpprc REQUEST=PESTPAIR ACTION=FAILBACK PPRCOPY ESTPAIR FAILBACK
  4. Failback from h1 to h3
    1. Issue a failback command to the h1 to h3 volumes.
      This process creates the original Multi-Target PPRC configuration as shown in Figure 4.
      TSO/E command DS CLI command ANTRQST API command ICKDSF command
      CESTPAIR ACTION(FAILBACK) failbackpprc REQUEST=PESTPAIR ACTION=FAILBACK PPRCOPY ESTPAIR FAILBACK
      Figure 4. Configuration after a failback from h1 to h2 and resuming production from h1 to h3
      The graphic a configuration after a failback from h1 to h2 and resuming production from h1 to h3