What steps will reproduce the problem?
1. run a heap of jobs that complete (this is a rare race condition)
What is the expected output? What do you see instead?
They should all complete. occaisonally one 'sticks' in exec:done
Please use labels and text to provide additional information.
What's happening from debug is once in a blue moon, the thread handling the
status POST request blocks in update_task_status at the task.save() line and
the thread sleeps. The other status results continue to be processed to
completetion, then finally the task unsleeps, and the exec:done status is
save()d and over writes the complete status.
Original issue reported on code.google.com by retrogra...@gmail.com on 19 Jul 2012 at 6:57
Original issue reported on code.google.com by
retrogra...@gmail.com
on 19 Jul 2012 at 6:57