Setting Up Extended Recovery for a IBM Connect:Direct/Plex Environment

Edit online

About this task

This section describes how to configure the IBM® Connect:Direct®/Plex environment to use extended recovery.

Procedure

  1. Add the following parameter to the initialization parameters file. 
    EXTENDED.RECOVERY=YES
  2. Add the XRFJOB DD statement to the IBM Connect:Direct startup JCL as follows: 
    //XRFJOB DD DISP=SHR,DSN=$CD.PLEX.JCL(CDMGRX)

    $CD.PLEX.JCL is the PDS where the active (current) IBM Connect:Direct startup JCL is located. The CDMGRX member is the standby IBM Connect:Direct startup JCL or command to run as a started task, which will be created later.

  3. To run as a started task, use one of the following, or skip to step 4 if you are not running as a started task.
    • To run the IBM Connect:Direct standby image as a started task on the same z/OS image as the IBM Connect:Direct active image, use the following as the first statement in the JCL. 
      START=membername,parms

      In this example, START=indicates to issue a START command. When the START command is issued, the equal sign (=) is replaced with a blank, and the entire statement is passed to z/OS as a command.

      For example, if START=HOSTJCL,X is the first statement, then START HOSTJCL,X is issued to IBM Connect:Direct for z/OS®.

    • To run the IBM Connect:Direct standby image as a started task on a different z/OS image in the sysplex as the IBM Connect:Direct active image, use the following as the first statement in the JCL. 
      /*$VS,‘command'

      Where command is the command you want JES to send to z/OS (in this case, a JES2 environment). Because the statement does not begin with “START=”, IBM Connect:Direct submits the statement to JES. JES identifies the /*$VS and issues the command to z/OS rather then placing it in the job queue.

      For example, if /*$VS,‘RO CSGB,S CDICOMB' is submitted as JCL, the

      RO CSGB,S CDICOMB command is issued rather than placed in the job queue.

  4. Create the CDMGRX member from the member that contains your MANAGER's JCL..
  5. Change the job name in CDMGRX so that it runs simultaneously with other IBM Connect:Direct/Plex members.
  6. Change the XRFJOB DD statement in CDMGRX to point to the CDMGR member, as in the following example. 
    //XRFJOB DD DISP=SHR,DSN=$CD.PLEX.JCL(CDMGR)
  7. Add the following DD statement to the CDSRV1 member (the SERVER1 startup JCL) in $CD.PLEX.JCL. 
    //XRFJOB DD DISP=SHR,DSN=$CD.PLEX.JCL(CDSRV1X)
  8. Copy the changed CDSRV1 JCL member to CDSRV1X.
  9. Change the job name in CDSRV1X so that it runs simultaneously with other IBM Connect:Direct/Plex members.
  10. Change the XRFJOB DD statement in CDSRV1X to point to CDSRV1, as in the following example. 
    //XRFJOB DD DISP=SHR,DSN=$CD.PLEX.JCL(CDSRV1)
  11. Add the following DD statement to the CDSRV2 member (the SERVER2 startup JCL) in $CD.PLEX.JCL. 
    //XRFJOB DD DISP=SHR,DSN=$CD.PLEX.JCL(CDSRV2X)
  12. Copy the changed CDSRV2 member to CDSRV2X.
  13. Change the job name in CDSRV2X so it can run simultaneously with other IBM Connect:Direct/Plex members.
  14. Change the XRFJOB DD statement in CDSRV2X to point to CDSRV2, as in the following example. 
    //XRFJOB DD DISP=SHR,DSN=$CD.PLEX.JCL(CDSRV2)
  15. Submit the CDMGR JCL to bring up the IBM Connect:Direct image using extended recovery.

    Each IBM Connect:Direct member initializes using extended recovery and submits the JCL specified in its XRFJOB DD statement. This JCL starts the standby IBM Connect:Direct members. Each standby IBM Connect:Direct member partially initializes, then begins monitoring its active IBM Connect:Direct member.

    If the active member terminates abnormally, or is shut down with a STOP CD CDPLEX RECOVER command, the standby member resumes initialization, becomes the active member, and submits the JCL in its XRFJOB DD statement. This JCL initializes the original active IBM Connect:Direct/Plex member, which now becomes the standby member.

    Note: If you want the standby member to run on a different z/OS image in the sysplex, you must define the VTAM APPLID as dynamic in both z/OS images, and you must define TCP/IP addresses as dynamic VIPA addresses.