Automatic Restart Management Support

Automatic restart management can reduce the impact of an unexpected error to a system program by restarting it automatically, without operator intervention. In a Sysplex environment, a system program can enhance its own recovery potential by registering as an element of the automatic restart management function of the cross system coupling facility (XCF).

To provide program recovery through automatic restart management, your installation must activate a policy through the SETXCF START command. This can be an installation-written policy or part of the IBM-supplied policy defaults. Because an installation-written policy may affect your program, you must understand how your installation uses automatic restart management for recovery in a Sysplex. Details of the automatic restart management function are described in the section of IBM manual MVS Programming: Sysplex Services Guide that discusses using the automatic restart management function of XCF.

In general, the operating system restarts an element under the following conditions:

• when the element itself fails—in this case, the operating system restarts the element on the same system
• when the system on which the element was running unexpectedly fails or leaves the Sysplex—in this case, the operating system restarts the element on another system in the Sysplex. This is called a cross-system restart.

The operating system will not attempt to restart an element if

• the element has not yet registered as an element of automatic restart management
• the element is cancelled through a CANCEL, STOP, or FORCE command (without the ARMRESTART parameter specified)
• JES is down or has indicated that the element should not be restarted
• the element has reached the restart attempts threshold specified in the policy
• the policy indicates that this element should not to be restarted
• access to the ARM couple data set was lost