The Redis broker uses a non-zero value for max_retries to determine if it should remove a job from the "running" state. So when an AbortException is thrown, max out the job.retries counter, rather than zeroing out the job.max_retries counter. This will have the same effect, but play more nicely with what the Redis broker is expecting.
The Redis broker uses a non-zero value for
max_retries
to determine if it should remove a job from the "running" state. So when anAbortException
is thrown, max out thejob.retries
counter, rather than zeroing out thejob.max_retries
counter. This will have the same effect, but play more nicely with what the Redis broker is expecting.The MemoryBroker is unaffected by this.
Fixes Issue #20