Wednesday, 2021-09-22

-opendevstatus- NOTICE: Zuul has been restarted in order to address a performance regression related to event processing; any changes pushed or approved between roughly 17:00 and 18:30 UTC should be rechecked if they're not already enqueued according to the Zuul status page18:37
timburkealmost meeting time!20:56
timburke(i need to update the agenda...)20:56
timburke#startmeeting swift21:00
opendevmeetMeeting started Wed Sep 22 21:00:15 2021 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.21:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.21:00
opendevmeetThe meeting name has been set to 'swift'21:00
timburkewho's here for the swift meeting?21:00
kotagood morning21:04
timburkeo/21:04
kotatimburke: o/21:04
timburkelooks like it might be a pretty short meeting ;-)21:04
timburkeas usual, the agenda's at https://wiki.openstack.org/wiki/Meetings/Swift21:05
timburke#topic eventlet / 100-continue bug21:05
timburkejust a heads-up, since i think we mentioned this last week21:06
timburkehttps://github.com/eventlet/eventlet/pull/578 merged21:07
timburkethis addresses some 400s our clients would occasionally see when re-using a connection where the client sent Expect: 100-continue but then we didn't send a 100 Continue and instead delivered a final response immediately21:07
timburkethe trouble was we'd see clients try *either* send a body for the request that was attempted *or* start right in on a new request, so the only option left to us to get on a sane path was to close the connection21:09
timburke#topic ring v221:09
timburkei added some tests for most of the patches on my chain, but looks like i left a line too long fairly early in the chain. expect a new set of patches coming, but it shouldn't materially change. i think the core of the format changes are ready for review21:11
timburke#topic root shard epoch reset21:11
timburkemy understanding of the current state is that we're planning to try to prevent None epochs from propogating21:11
timburkewe'll use that to see if we can track down the node that brought it on and from there figure out the root cause of how we got there, and address *that*21:12
timburkei was going to confirm with mattoliver that my understanding was correct, but i'm *pretty sure* that's the current plan21:12
timburkeso we (NVIDIA) might carry https://review.opendev.org/c/openstack/swift/+/809969 (or something close to it) for a bit until we suss out where these None epochs are coming from21:13
timburke#topic misplaced objects in sharded containers21:15
timburkei'm afraid i still haven't been following along on this enough to feel informed on the state of things21:15
timburkei was going to check if the updates on https://review.opendev.org/c/openstack/swift/+/805389 and https://review.opendev.org/c/openstack/swift/+/809317 were indicative of some progress; if nothing else, i suppose it's staying fresh in mattoliver and acoles's minds21:16
timburkethat's all i had to bring up21:17
timburke#topic open discussion21:17
timburkeall right, i think i'm'a call it21:19
timburkethanks for coming, kota -- make me feel like i'm not just talking to myself :-)21:19
kotathanks timburke for sharing current status21:19
timburkethank you for coming, and thank you for keeping up on swift!21:20
timburke#endmeeting21:20
opendevmeetMeeting ended Wed Sep 22 21:20:12 2021 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)21:20
opendevmeetMinutes:        https://meetings.opendev.org/meetings/swift/2021/swift.2021-09-22-21.00.html21:20
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/swift/2021/swift.2021-09-22-21.00.txt21:20
opendevmeetLog:            https://meetings.opendev.org/meetings/swift/2021/swift.2021-09-22-21.00.log.html21:20

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!