Race conditions involving submission processing and cancellation tasks

[jennifer-richards]

Describe the issue

Following a server restart, the celery workers were not running while a couple of submissions came in. When this was noticed and the celery worker started, there were a couple problems indicating races. First, for some reason each submission was canceled 4 (or so) times according to its log. Second, in one case, after the cancellation, the submission was marked as having passed its validation checks. This indicated that it was in the process of being processed when it was canceled.

It'd be nice to understand and avoid the multiple cancellations.

We need to be sure a canceled submission is not later marked as processed or otherwise allowed to continue its lifecycle.

Code of Conduct

• [X] I agree to follow the IETF's Code of Conduct