After a fatal death of the beat process, lock "celery:beat:task_lock" will block forever, every new process, resulting:
"another beat is running, disable this node util it is shutdown"
This issue requires a manual action (DEL "celery:beat:task_lock").
In my opinion, the lock should be set with a TTL (like 2*60s), and is should be upgraded for another TTL periodically (like every 60s), while process is alive. In case of an unexpected termination there is a chance for another beat to take a baton, after a while.
After a fatal death of the beat process, lock "celery:beat:task_lock" will block forever, every new process, resulting:
This issue requires a manual action (DEL "celery:beat:task_lock").
In my opinion, the lock should be set with a TTL (like 2*60s), and is should be upgraded for another TTL periodically (like every 60s), while process is alive. In case of an unexpected termination there is a chance for another beat to take a baton, after a while.
Regard