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.

If the root activity is activated by the process timer, it could, for example:
  1. 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.
  2. 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.
  3. 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.


dfhp9c0083.html | Timestamp icon Last updated: Monday, 5 June 2017