Resuming operations at the recovery site

If PPRC volumes are participating in a Global Mirror session, refer to Failover/failback for a Global Mirror session for recovery procedures. Otherwise, perform the following steps to have the recovery site assume control of operations:

  1. Reconfigure ESCON® Directors, if necessary, to connect local systems to recovery system Disk.
  2. Query each secondary disk volume to verify that it was a secondary remote copy volume and that it is in duplex state. Continue to the steps for volume states other than duplex state.
  3. If the volume is in duplex or suspended state, check the console for SYSLOG error messages that indicate that the ERP has attempted to put the volume in suspended state. If there is an error message, recover the volume as if it was in the suspended state at the time that the IEA49xx message indicates.
  4. Issue the PPRC CRECOVER command to change the secondary volume from its current state (duplex, suspended, or pending) to simplex state in order to let that volume come online. Table 1 shows suggested peer-to-peer remote copy volume recovery procedures that are based on the state of the volume at the time of system failure.

    Rules: The PPRC recovery function must have access to each secondary volume on all recovery site storage controls at the time of the recovery.

    Some recovery system volumes may not initially be usable for recovery. Change these volume serial numbers to prevent any applications from accessing the volumes until you can either restore them, or otherwise validate their contents by recovery procedures.

    Table 1. Volume failure states and recovery actions
    If volume is in . . .    Then do this . . .
    Duplex state Issue the CRECOVER command, as in the following example: 
    CRECOVER DEVN(X'0F40') PRIM(X'6060' 62019 X'00')
            SEC(X'6061' 68006 X'09') ID(DATA01)
    This returns the volume to simplex state, fully recovered and ready to use. (Issue the CQUERY command again if you wish to ensure that the volume is in simplex state.) You can now vary the volume (DATA01 in the above example) online for system use. Review the steps of resuming operations at the recovery site.
    Pending state If the volume is in pending state, then it never reached duplex state. Issue the CRECOVER command, as in the following example: 
       CRECOVER DEVN(X'0F40') PRIM(X'6060' 62019 X'00')
            SEC(X'6061' 68006 X'09') ID(DATA01 DATA02)
    This returns the volume to simplex state and changes the volume serial number from its old value to a new value. (Issue the CQUERY command again if you wish to ensure that the volume is in simplex state.) You can now vary the volume (DATA02 in the above example) online to do normal recovery from backup tapes. Once the volume state is fully recovered, the volume can be changed back to its original volume serial number so that applications can allocate and use the volume.
    Simplex state Determine why the volume is in simplex state before you accept the secondary copy as current and usable. The simplex volume may either not be a remote copy volume, or the volume may be in simplex state following a previous action that ended the volume copy pair. When you are satisfied with the volume status, vary the volume online for system use.
    Suspended state Determine what caused this state to occur. If the volume was suspended by the host or primary site storage control, data on the volume is valid. How current the data is depends on the timestamp information provided in the SYSLOG IEA49xx message. If the volume was suspended by the recovery site storage control, data on the secondary volume may be downlevel or incorrect for various reasons (i.e., an equipment check). In this case, recover from the backup data volumes.
    If the data on the volume is valid, issue the CRECOVER command, as in the following example: 
       CRECOVER DEVN(X'0F40') PRIM(X'6060' 62019 X'00')
            SEC(X'6061' 68006 X'09') ID(DATA01 DATA02)
    This returns the volume to simplex state and changes the volume serial number from its old value to a new value. (Issue the CQUERY command again if you wish to ensure that the volume is in simplex state.)

    With the volume (DATA02 in the above example) now in simplex state, vary the volume online and use the system log, database logs, and backup data volumes (when appropriate) to complete the recovery of data from the time the volume went into suspended state until the time of failure. Once the volume state is fully recovered, the volume can be changed back to its original volume serial number so that applications can allocate and use the volume.
  5. Restart the application systems and perform the same system and application startup procedures that you perform on the primary system when you restart there following a system failure.

    Recovery system catalog entries will be consistent with those on the primary system if catalog volumes are part of the data that was copied to the recovery system. See Copying the catalog and control data sets for guidelines on managing catalog updates that are not copied to the recovery system.

    If the catalog change rate is very low, it may be more effective to copy the catalog with remote copy. Consider what other data sets are on the volumes with the catalogs, as PPRC will also copy these data sets to the recovery system.

Parent topic: Peer-to-Peer Remote Copy data recovery operations