Hang up for a long time and enter failed in about half an hour
tinnkai opened this issue · 4 comments
@tinnkai thank you for opening an issue!
Would you mind showing the output of
redis-cli ZRANGE asynq:{default}:deadlines 0 -1 WITHSCORES
@hibiken
It may be caused by the direct shutdown of the service while the queue is being processed,But what I expect can I keep the failure queue and then I can do it manually.
thank you
@hibiken
Sorry, I misunderstood. Processed includes the number of failed and successful, that is, entering failed will retry
@tinnkai Thanks for providing additional info!
I think initial confusion around task being stuck in "active" state is due to how the task recovery works currently when a worker crashes (or forced killed kill -9
). Currently, asynq recovers task if its deadline has expired over 30s ago, but we have an open issue to improve this by recovering these orphaned tasks more quickly (hibiken/asynq#334).