In short, see https://github.com/petemoore/test-supersede/ This will create a set of 5 tasks, where one supercedes all the others. The superceding task intentionally fails (max run time exceeded). All of the tasks that should be superseded by it are then run individually (with no superseding) since superseding only works for task run 0 (which has been "used up" with a fake "worker-shutdown" exception in the abort of the primary task). The test-supersede repo above demonstrates the problem.
That's intentional: we don't want to try to repeatedly supersede a task, so we only supersede run 0. Re-runs are fairly rare, so this is a nice way of "failing open"
Status: NEW → RESOLVED
Last Resolved: a year ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.