Steps for recovering from processor, address space, and XRC data set failures

XRC error conditions may occur as a result of various failures on an application processor. These errors can cancel critical address spaces that reside within the affected system, and may in turn cause the application to end. The system data mover may also be on the system with the failure.

Other error situations may affect only the SDM system. The SDM may be affected by an environmental problem, an internal SDM error, errors on the XRC state, control, or journal data sets, or due to the operator canceling the ANTASnnn address space.

Before you begin: You need to be familiar with the commands used in these steps. For more information, refer to Extended remote copy command descriptions.

Perform the following steps to recover from processor, address space, and XRC data set failures:

  1. Correct the error.
  2. Restart the primary system, if necessary.
  3. Restart the SDM system, if necessary. For many errors, the XRC session will restart itself.
  4. Issue an XSTART command to start the XRC session, if necessary.
  5. Issue an XADDPAIR SUSPENDED command to add the suspended volume pairs back to the XRC session. XRC automatically copies all pending updates when the application programs become active.
  6. Issue the XCOUPLE msessionid RELEASE command, if the XRC session is coupled. To clear up a cluster session, issue the MVS MODIFY command of F ANTAS000,CREFRESH FORCE. See Advanced Copy Services diagnostic aids for a description of this command. You can use the F ANTAS000,CREFRESH FORCE command in the following circumstances:
    • The ANTCLnnn address space is not active, and you cannot enable the ANTCLnnn address space by issuing the CREFRESH command without FORCE.
    • The ANTCLnnn address space is active, and you cannot deactivate the ANTCLnnn address space by issuing the CREFRESH command without FORCE.
    • The master session consistency time does not advance, and messages identify the cluster as the reason that the master consistency time is delayed.

    In a disaster environment, you might impact data consistency if you incorrectly use the MVS MODIFY command of F ANTAS000,CREFRESH FORCE.