Considerations for jobs that can affect error recovery

When a line or controller fails, and the application programs are notified, often you must end those jobs that were running over the line and controller. Those jobs must be started again after the communications resource is recovered. Jobs ending (particularly abnormal job ending) should be considered from a performance perspective as an extremely complex transaction. Use the following links to help you recover from an abnormal job ending.

• Device recovery, see Considerations for the CMNRCYLMT parameter that can affect error recovery
• Prestart jobs, see Change prestart job entries that can affect APPC error recovery
• Joblog generation, see Joblog considerations that can affect communications error recovery
• Using the Change System Job (CHGSYSJOB) commands

  The CHGSYSJOB command allows you to change the run priority of a system job. The following are system jobs of interest for communications recovery:

  • QCMNARB01 through QCMNARB99
  • QSYSCOMM1

  In general, these system jobs should be allowed to run at their default, system-provided priority. However, if one of these jobs begins using large amounts of CPU, and is affecting other work on the system, it is possible to lower its priority. Note that this may result in queued-up work for that job.

• Device Wait Timeout

  The device wait (DEVWAIT) timeout is used to limit the amount of time a subsystem waits for workstation input/output to complete.