Using process timers
Use process timers to restart a stuck process. You can set timers for individual child activities, for the process, or for both. That is, the root activity can set a timer with an expiry time after the whole process might be expected to complete.
About this task
If the process is short-lived, you might decide not to set any activity timers, but to set a process timer instead.
If the process is long-lived, do not set a process timer without also setting timers for at least some individual activities. Setting timers prevents the possibility of a delay in restarting the process. For example, if a process that is expected to last six months becomes stuck after one day while processing its first activity, and you have set only a process timer, the process could lie dormant for, say, seven months before the root activity is reactivated to deal with the problem.
- Browse and inquire on each of its descendant activities, checking completion status and mode. For examples of the use of the BTS browsing and inquiry commands, see Browsing examples.
- If it succeeds in identifying the stuck activity, issue a CANCEL command to cancel it. If the stuck activity is not a child but a lower-level descendant of the root activity, the root must first acquire the stuck activity.
- The completion event of the stuck activity fires, causing the parent activity to be reactivated. The CHECK ACTIVITY command issued by the parent returns a completion status of FORCED. The parent should be coded to handle the abnormal completion of one of its children. The process is no longer stuck.