I think I might have seen too, but that was a long time ago...
I can't reproduce this in my lab setting. Here's what I did:
- configured object-replicator as following:
[object-replicator] sync_method = rsync rsync_timeout = 60 lockup_timeout = 5 stats_interval = 2
- uploaded some multiple GB-sized objects - deleted all but one replica of these objects, and hashes.pkl too - replicator runs, I see the "killing live coros" - and it continues?
Not sure how to reproduce this yet.
Looking at the code I think there is another possibility to run into this in case the top-level replicator loop raises a Timeout (https://github.com/openstack/swift/blob/master/swift/obj/replicator.py#L759-L760).
I think I might have seen too, but that was a long time ago...
I can't reproduce this in my lab setting. Here's what I did:
- configured object-replicator as following:
[object- replicator]
sync_method = rsync
rsync_timeout = 60
lockup_timeout = 5
stats_interval = 2
- uploaded some multiple GB-sized objects
- deleted all but one replica of these objects, and hashes.pkl too
- replicator runs, I see the "killing live coros" - and it continues?
Not sure how to reproduce this yet.
Looking at the code I think there is another possibility to run into this in case the top-level replicator loop raises a Timeout (https:/ /github. com/openstack/ swift/blob/ master/ swift/obj/ replicator. py#L759- L760).