21:00:13 <timburke> #startmeeting swift
21:00:13 <opendevmeet> Meeting started Wed Feb 16 21:00:13 2022 UTC and is due to finish in 60 minutes.  The chair is timburke. Information about MeetBot at http://wiki.debian.org/MeetBot.
21:00:13 <opendevmeet> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
21:00:13 <opendevmeet> The meeting name has been set to 'swift'
21:00:19 <timburke> who's here for the swift meeting?
21:00:41 <kota> hi
21:01:17 <mattoliver> o/
21:01:41 <acoles> o/
21:01:55 <zaitcev> o/
21:02:08 <timburke> as usual, the agenda's at https://wiki.openstack.org/wiki/Meetings/Swift
21:02:14 <timburke> #topic vPTG
21:02:35 <timburke> i made a doodle poll to figure out meeting times
21:02:39 <timburke> #link https://doodle.com/poll/qs2pysgyb8nb36c2
21:03:13 <timburke> we'll aim for 4-5 3-4hr sessions like in the past
21:03:37 <timburke> and we can pick a slot for another ops feedback session
21:03:39 <mattoliver> kk
21:04:20 <timburke> if you haven't already, don't forget to register
21:04:24 <timburke> #link https://openinfra-ptg.eventbrite.com/
21:05:34 <timburke> and i'll start an etherpad for us to gather topics over the next month and a half or so
21:05:54 <timburke> any questions on PTG logistics?
21:07:30 <acoles> I just did the doodle poll but can't see my entry after submitting so cannot check them
21:08:08 <timburke> i can see it :-) maybe i should have left responses public
21:08:26 <acoles> timburke: as long as you can see it and tell me when to show up that's all I need :)
21:08:49 <timburke> all right
21:08:57 <timburke> #topic Swift 2.29.0
21:09:14 <timburke> we had a release!
21:09:18 <acoles> yay!
21:09:25 <mattoliver> \o/
21:10:17 <kota> excellent
21:10:43 <timburke> #link https://docs.openstack.org/releasenotes/swift/current.html
21:10:53 <timburke> (sorry, i should have dug that out earlier :-)
21:11:39 <timburke> that's a lot of great improvements, across a lot of subsystems
21:13:13 <timburke> that's all i've got -- i've been so focused on the release and election stuff, i haven't looked hard at what our next tactical items should be
21:13:19 <timburke> #topic open discussion
21:13:41 <zaitcev> I don't have anything either.
21:14:03 <mattoliver> Thanks for all that work timburke totally understandable.
21:15:11 <mattoliver> I've been blowing the dust of the missing pieces of shard shrinking. But maybe more on that next week when I have something more concrete to talk about rather then reloading it in my head :)
21:15:58 <timburke> oh yeah, and i still need to look at your ring format patch...
21:16:17 <mattoliver> I pushed up a PTG version of the v2 serialisation stuff. If anyone wants to comment esp if I incorp all the things from the last PTG
21:16:30 <mattoliver> lol, yeah, that. (no rush though)
21:17:37 <mattoliver> I also have a chain starting: https://review.opendev.org/c/openstack/swift/+/828069 that was started by attempting to debug a replication problem but the entries in the sync tables where a dead end.
21:18:38 <timburke> cool
21:18:56 <mattoliver> ^ that makes the db ids more debug friendly so long as you have unique device names in your ring. The follow up adds the info to swift-{container,account}-info (optionally).
21:19:54 <mattoliver> *adds the ability to dump the sync tables to the cli info tool
21:19:56 <zaitcev> Unique as in not 10.0.0.4:sda and 10.0.0.5:sda?
21:20:19 <mattoliver> yeah. Otherwise it's about as useful as it currently is.
21:20:52 <mattoliver> although I guess limits it to 1/n of the cluster where n is the number of nodes :P
21:21:12 <mattoliver> no that math is wrong
21:21:14 <mattoliver> too early
21:21:17 <mattoliver> :P
21:21:22 <mattoliver> but you get the idea.
21:22:05 <timburke> i recently started playing with adding a delimiter-depth query param: https://review.opendev.org/c/openstack/swift/+/829605
21:22:19 <mattoliver> We number devices, like d<num> to each device is unique. so the above patch means I can just look in the ring to know exactly which node attempted to replicate at that time.
21:23:19 <timburke> we've got some users that laid out data like `AA/BB/AABBCCCCCCCCC/` and they want all of those AABBCCCCCCCCC values, but there's like 64k possible AABB combinations...
21:24:33 <timburke> all right -- let's call it early
21:24:46 <timburke> thank you all for coming, and thank you for working on swift!
21:24:51 <timburke> #endmeeting