Ceph » CephFS

Right now:

                    while True:
                        if self.should_reconfigure_num_threads():
                        log.info("thread [{0}] terminating due to reconfigure".format(thread_name))
                            self.async_job.threads.remove(self)
                            return
            vol_job = self.async_job.get_job()
            if vol_job:
                            break
            self.async_job.cv.wait()

If there is no work, then just wait - which is fine since this was written with the assumption that another thread would queue a "work" and then kick the async thread(s). But lately, we ran into a situation where the work had to be queued "manually" by dropping directories into `/volumes/_deleting`. This requires failing ceph-mgr. Adding a timeout argument to periodically check for available work would be beneficial.