Avoiding a single point of failure

You should never have a single point of failure in your XCF signaling configuration. Loss of a single CTC link or single coupling facility structure could cause total loss of signaling connectivity between two or more systems in the sysplex. To avoid a single point of failure in the sysplex, specify the following in the CFRM policy:
  • The names of at least two list structures for signaling
  • For each list structure you define, the names of at least two coupling facilities in the preference list
  • Exclusion lists, so that the list structures are allocated in different coupling facilities.

If you are especially concerned about availability, you can use CTC devices, as well as coupling facilities, for signaling because they provide a different transport mechanism with different hardware and software points of failure. However, this approach increases the complexity of systems management.

You also need to ensure that there is enough free space in the coupling facilities to allow MVS™ signaling services to recover from failures by rebuilding list structures. For high availability, you should never rebuild all the signaling structures in your XCF signaling configuration at the same time, unless there is adequate redundant CTC connectivity to compensate for the temporary unavailability of the signaling structures. While it is being rebuilt, an XCF signaling structure is quiesced and therefore is unusable until the rebuild process completes. Without redundant CTC connectivity during rebuild of XCF signaling structures, structure rebuild processing is significantly slowed down. This condition may result in one or more systems being removed from the sysplex when there is an active SFM policy with CONNFAIL(YES) specified.

Note that the REBUILDPERCENT parameter does not apply to signaling structures.