21:00:45 <timburke> #startmeeting swift
21:00:45 <opendevmeet> Meeting started Wed Aug  3 21:00:45 2022 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:45 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:45 <opendevmeet> The meeting name has been set to 'swift'
21:00:51 <timburke> who's here for the swift meeting?
21:01:08 <kota> hi
21:02:02 <mattoliver> o/
21:02:22 <timburke> as usual, the agenda's at https://wiki.openstack.org/wiki/Meetings/Swift
21:02:33 <timburke> #topic sha1 deprecations
21:02:58 <timburke> tempurl and formpost patches have now all merged, and i abandoned the client backport
21:03:20 <timburke> i think the only thing left is to do the same dance for /info admin calls
21:03:25 <timburke> #link https://review.opendev.org/c/openstack/swift/+/850787
21:03:51 <timburke> should take care of it, just needs review
21:04:17 <mattoliver> I'll take a look at that today
21:04:40 <timburke> thanks!
21:04:58 <timburke> #topic ring v2
21:05:11 <timburke> i finally goaded clayg into reviewing the head of the chain
21:05:14 <timburke> #link https://review.opendev.org/c/openstack/swift/+/834261
21:06:13 <timburke> i'll try to keep on him to make sure it's to his liking :-)
21:06:51 <mattoliver> Cool
21:06:52 <timburke> and i should also rebase some of the later patches in the chain
21:07:09 <clayg> i mean, if you think my feedback like "swift-ring-builder write_ring --format=v2 should write a v2 ring" is un-reasonable... ;)
21:07:56 <mattoliver> Lol, what kinda cli interface is that 😜
21:08:28 <mattoliver> You hot to do it while standing on your head.. obviously
21:08:41 <clayg> yeah, I think I still have a bunch of work to do to give a serious review - what are the parts you think need the most testing or validation?  any design choices you're still un-easy about?
21:09:05 <timburke> clayg, that was great! mattoliver and i have been so in the weeds on it (and through so many rebases and restarts) that we missed that kind of sniff test
21:09:34 <clayg> i remember I didn't understand some of the scary doc-strings in the internal bits of the ring.io module 🤔
21:10:11 <timburke> i did my best to wrap it up so no one else would *have* to understand it ;-)
21:10:46 <timburke> i suppose one thing i haven't spot-checked is deserialization speed -- i remember at some point in investigating the history of ring formats, that was part of what drove us away from pickles
21:13:16 <timburke> #topic backend rate-limiting
21:13:52 <timburke> i think we're still bogged down with other things and haven't gotten around to testing the patch in prod
21:14:22 <timburke> but wanted to give others a chance to correct me if i'm wrong :-)
21:14:27 <clayg> maybe we could prioritize that - since we're waiting on new rev of the no-lock-logging patch
21:15:42 <timburke> i already pushed up a new one for that -- and i'm going to take a look at making the related eventlet toggle per-fileno instead of global
21:18:06 <timburke> (long and short of that is we tried https://review.opendev.org/c/openstack/swift/+/840232 on some real hardware, which then started tripping https://github.com/eventlet/eventlet/blob/master/eventlet/hubs/hub.py#L176 since we were no longer suppressing it at https://github.com/openstack/swift/blob/master/swift/common/utils.py#L6196 )
21:19:32 <timburke> though i like prioritizing the backend ratelimiting patch, too -- and it seems like we could probably re-use some of the workload generation for testing both patches
21:20:50 <timburke> all right -- one more last-minute topic
21:21:00 <timburke> #topic 2.30.0 release
21:21:41 <timburke> i'm getting some release notes together! i wanted to do this last week, but didn't get around to it. hopefully by friday there'll be a changelog patch up
21:21:56 <timburke> that's all i've got
21:22:01 <timburke> #topic open discussion
21:22:09 <timburke> what else should we talk about this week?
21:23:01 <zaitcev> I've got nothing.
21:23:29 <clayg> @timburke i don't understand why the no-lock-logging setup triggers the warning tho - who's waiting on to read what now?
21:24:38 <timburke> clayg, it was complaining about *writes*, not reads. the escape hatch is poorly named
21:25:17 <timburke> the trouble was that we had multiple greenthreads trying to write to the logging socket at the same time
21:26:38 <timburke> which is exactly what the patch expected to have happen; the theory is that because of us using UDP, that should be fine
21:27:09 <clayg> really?  at the *same* time?!  can you really get a /dev/log to raise EAGAIN... maybe...
21:28:48 <timburke> i was surprised, too -- but i still think it's going to be handled as intended
21:29:09 <timburke> i'm going to be out camping next week -- would anyone like to chair the meeting, or should we just cancel and meet the following week?
21:30:03 <mattoliver> Depends if we have anything to talk about. How about we tentively cancel. And I'll chair and let people know if anything changes :)
21:30:23 <timburke> 👍
21:30:28 <mattoliver> But I suspect canceling is fine
21:30:39 <mattoliver> Enjoy your camping :)
21:31:10 <kota> enjoy
21:31:18 <clayg> 🏕️
21:31:51 <timburke> all right, i think i'll call it early then
21:32:04 <timburke> thank you all for coming, and thank you for working on swift!
21:32:09 <timburke> #endmeeting