Wednesday, 2026-02-04

opendevreviewMerged openstack/swift feature/threaded: Merge remote-tracking branch 'gerrit/master' into feature/threaded  https://review.opendev.org/c/openstack/swift/+/97557501:05
opendevreviewWael Halbawi proposed openstack/swift master: relinker: Prefix log messages with invoked action  https://review.opendev.org/c/openstack/swift/+/97555205:16
opendevreviewTim Burke proposed openstack/swift feature/mpu: Merge remote-tracking branch 'gerrit/master' into feature/mpu  https://review.opendev.org/c/openstack/swift/+/97559105:38
timburke#startmeeting swift21:00
opendevmeetMeeting started Wed Feb  4 21:00:03 2026 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 team meeting?21:00
mattolivero/21:00
timburkei only just updated the agenda21:01
timburke#link https://wiki.openstack.org/wiki/Meetings/Swift21:01
timburkefirst up, a few broader openstack announcements21:02
timburke#topic 2026.2 cycle21:02
timburkewe've got an official codename: following gazpacho will be hibiscus!21:03
timburke#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/TGTZO53DLCVR7PQT37VHSIXL6EAYHMYO/21:03
mattoliverHibiscus.. ok21:03
timburkefwiw i think it was my favorite of the options (though i don't actually remember what the others were now...)21:04
timburke#topic vPTG21:04
timburkewe've also got dates for April's vPTG! it'll be Apr 20-2421:04
timburke#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/65JHXUYJJUPSAXPY7AJFR2VYWNZUJTWT/21:04
timburkeand registration is already open21:04
timburke#link https://ptg.openinfra.org/21:05
mattoliverOh another ptg! We should land some things, so we don't have to talk about them again :p21:05
timburkei still need to respond to the team survey, set up an etherpad, etc21:05
timburke#topic elections21:05
mattoliverLet me know if you want me to help organise anything Tim21:06
timburkeand as an fyi, election nominations (for both TC and PTL) will be open shortly21:06
timburkethanks mattoliver -- i might take you up on that as we get closer21:07
mattoliverNps21:07
timburkefwiw, i plan on running again -- definitely have a lot more time for upstream work (and work in general) compared to a couple months ago21:08
timburkenext up21:08
timburke#topic releases21:08
mattoliverWoo! Yeah glad to hear it21:08
timburkei got some releases out! not only a 2.37.0 cut from master, but also fresh stable tags for all open branches21:09
mattoliverOh yeah saw a bunch of those release notes patches21:09
mattoliverNice one21:09
timburkei should probably look at the state of python-swiftclient... but i don't think there are any new patches landed since the last release21:11
timburkeoh yeah, i forgot i'd written p 961423 ...21:12
patch-bothttps://review.opendev.org/c/openstack/python-swiftclient/+/961423 - python-swiftclient - authv1: Modernize returned service catalog - 3 patch sets21:12
timburkenext up21:12
timburke#topic docker images21:13
timburkethe swift image upload/promote jobs are fixed! thanks mattoliver, clarkb, and fungi for the help in figuring that out21:13
mattoliverYay \o/21:14
fungiglad it got working!21:14
timburkei also realized (and then figured out that mattoliver had *also* realized) that the promote jobs also run on the stable branches... so "latest" occasionally jumps around between master and stable...21:15
mattoliverOuch21:16
mattoliverIt is a little nuts21:16
timburkei'm trying to figure out how best to fix that -- i think what i really want is some dynamism to the set of tags to promote to, 'cause i'd also really like to get per-release docker tags21:16
fungimaybe don't call any of them "latest" and instead tag them with the originating branch names21:17
fungithe branch name should be in a job variable accessible from the ansible playbook21:17
mattoliverYeah and only master is latest or something21:18
mattoliverIt should be some ansible fu21:18
fungisure, if you wanted that one to continue to be called "latest" then a simple conditional check for ==master or whatever21:18
timburkeyeah, that's roughly the approach mattoliver took in p 732480 -- and yeah, i think we could get some ternary op to relabel master to latest21:19
patch-bothttps://review.opendev.org/c/openstack/swift/+/732480 - swift - Add branch and tag tags to docker image - 3 patch sets21:19
mattoliverYeah something like that!21:19
fungithere are also string mangling operators if you wanted to strip a "stable/" prefix form the tag name21:19
fungier, from21:19
timburkebut i'm not sure the "{{ zuul.tag | default('') }}" is going to work well -- i didn't see any check for non-empty tag names in https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/promote-docker-image/tasks/promote-retag.yaml21:21
timburke(maybe i should propose it?)21:21
mattoliverMaybe I'll dig back into that maybe. 5 year ago me is a different Matt, so need to figure out what I was thinking :p21:22
timburkeyeah, i was thinking of replacing / with - (since / won't be allowed in tag names)21:22
mattoliverYeah dash is probably a better option21:22
timburkewe'll play around with it, anyway21:24
timburkenext up21:24
timburke#topic timestamp collisions21:24
timburkea bunch of us met up on a video call yesterday to talk about the changes -- i think we're making some decent progress21:25
mattoliver+121:25
timburkeone thing i realized afterwards was that there's a lot of interplay between what's going on with timestamps and what we're trying to do on feature/mpu21:26
timburkei tried merging from master last night... it was a bit of a bear, and i still don't have all the tests passing21:26
timburke#link https://review.opendev.org/c/openstack/swift/+/97559121:27
patch-botpatch 975591 - swift (feature/mpu) - Merge remote-tracking branch 'gerrit/master' into ... - 1 patch set21:27
mattoliverWoah, that'll be a big rebase, we've landed 44 pre patches for ts collisions21:27
mattoliverAnd alot of test churn21:27
timburkei'm realizing that i definitely want fresh merges from master to feature/mpu both just before and just after we start introducing the jitter (or whatever we end up calling it) to all uploads21:28
timburkeyeah, the changes in container/test_backend.py were a pain to stitch up21:30
mattoliverYeah fair enough21:30
mattoliverWell nice job with the master merge, it will need to happen, so thanks for taking it on21:32
timburkebut i might be more nervous about whether i got the proxy changes right -- between p 913832 and p 956411 things got messy21:32
patch-bothttps://review.opendev.org/c/openstack/swift/+/913832 - swift (feature/mpu) - proxy: Add list of backend response data to reques... (MERGED) - 10 patch sets21:32
patch-bothttps://review.opendev.org/c/openstack/swift/+/956411 - swift - Return 503 for POST request with mixed 202/404 res... (MERGED) - 7 patch sets21:32
timburkei'll keep chipping away at it -- we'll get there eventually21:33
timburke#topic LRC scheme for liberasurecode21:33
timburke#link https://review.opendev.org/c/openstack/liberasurecode/+/95928021:33
patch-botpatch 959280 - liberasurecode - feature: LRC (locally repairable code backend) - 17 patch sets21:33
timburkei think the patch is getting really close!21:33
mattoliverNice21:33
mattoliverI have a tab open on LRCs on my laptop to help me understand them, I should get back to reading it :)21:34
timburkei think i'm mostly trying to decide right now how hard to push back that the LRC-specific stuff should be in isa_l_rs_lrc.c instead of isa_l_common.c21:35
timburkeif anyone else has time to take a look at the code, i'd appreciate it21:35
mattoliverAhh ok, I'll take a look today21:35
mattoliverI can at least have an opinion on that21:35
timburkethanks21:36
timburkenext up21:36
timburke#topic libec release21:36
timburkeespecially if we get the LRC backend merged, i'd like to get a release up soon21:37
mattoliverYeah, that totally deserves a release21:38
timburkecould even do a 1.7.2 before then, for the improved locking21:38
mattoliverYeah that's true. Thats already a big change21:39
timburkebut i want to do *something* so i can start playing around with free-threaded python in pyeclib without making up local-only releases ;-)21:39
mattoliverAnd probably worth a release21:39
timburkesee also21:39
timburke#link https://review.opendev.org/c/openstack/pyeclib/+/97426321:39
patch-botpatch 974263 - pyeclib - Turn off GIL for free-threaded builds - 2 patch sets21:39
mattoliverWoo, love that one21:40
timburkeremarkably small and direct change :-)21:40
mattoliverWell I'm ok if you want to do 2 releases, lock and then LRC. But maybe give LRC a few more days of you think it's close21:41
timburkeyeah, i was planning on waiting a bit and just going straight to a 1.8.0 if lrc lands21:42
mattoliver+121:43
timburke#topic pyeclib binary wheel targets21:43
timburkei *also* wanted to start playing with publishing free-threaded wheels -- but i realized we'd need to update our base images for the builds21:44
timburke#link https://review.opendev.org/c/openstack/pyeclib/+/97545821:44
patch-botpatch 975458 - pyeclib - Update docker base images to supported versions - 1 patch set21:44
timburkewe'd lose manylinux1 and manylinux2010, but manylinux2014 seems like it ought to support most (all?) of the platforms we care about at this point21:45
mattoliverYeah, and whose such old bases nowadays21:46
mattoliver*Who uses21:46
timburkebut at least we've got confirmation that we can build free-threaded wheels in21:48
timburke#link https://review.opendev.org/c/openstack/pyeclib/+/97545921:48
patch-botpatch 975459 - pyeclib - Build free-threaded wheels - 1 patch set21:48
mattoliverCool21:48
timburkethough i think i want to wait on merging that until we've got a new libec to include in the wheel :-)21:48
timburke(and maybe also a new isa-l? looks like they're gearing up for a release)21:49
timburke#link https://github.com/intel/isa-l/issues/38421:49
timburkeall right, i think that's all i've got21:50
timburke#topic open discussion21:50
timburkeanything else we should bring up this week?21:50
mattoliverThen let's do a release21:50
mattoliverI've got nothing today21:50
timburkeall right, let's call it then21:52
timburkethanks for coming, and thank you for working on swift!21:52
timburke#endmeeting21:52
opendevmeetMeeting ended Wed Feb  4 21:52:25 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)21:52
opendevmeetMinutes:        https://meetings.opendev.org/meetings/swift/2026/swift.2026-02-04-21.00.html21:52
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/swift/2026/swift.2026-02-04-21.00.txt21:52
opendevmeetLog:            https://meetings.opendev.org/meetings/swift/2026/swift.2026-02-04-21.00.log.html21:52
mattoliverThanks [@_oftc_timburke:matrix.org](https://matrix.to/#/@_oftc_timburke:matrix.org) 21:52

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!