Thursday, 2025-10-23

*** mhen_ is now known as mhen01:45
tonybcroelandt: Are https://bugs.launchpad.net/glance/+bug/2121144 and https://review.opendev.org/c/openstack/requirements/+/959560 still relevant ?10:11
croelandttonyb: yes, as we have not yet provided a fix13:28
croelandttonyb: I should probably rebase the patch then, thanks for pointing that out13:28
croelandttonyb: done14:00
croelandt#startmeeting glance14:01
opendevmeetMeeting started Thu Oct 23 14:01:04 2025 UTC and is due to finish in 60 minutes.  The chair is croelandt. Information about MeetBot at http://wiki.debian.org/MeetBot.14:01
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:01
opendevmeetThe meeting name has been set to 'glance'14:01
croelandt#topic roll call14:01
croelandt#link https://etherpad.openstack.org/p/glance-team-meeting-agenda14:01
rosmaitao/14:02
mheno/14:02
dansmitho/ (but multistasking)14:02
croelandt#topic Release/periodic job updates14:02
croelandtALl good \o/ Except the fips one we know about14:02
croelandt#topic Important stable patches14:03
croelandt#topic https://review.opendev.org/c/openstack/glance_store/+/96187014:03
croelandtThere is a weird install issue for this one. This is annoying because it prevents us from merging the same patch in stable/2024.1 and the deadline is tomorrow :-(14:03
croelandtI think I should probably ask someone, but not sure who can help with these kind of install issues14:03
croelandtWe're trying to install both glance-store===4.8.2 from the upper constraints, and 4.8.3dev1 from... the repo itself, I guess14:04
croelandtIf anyone has input on how to fix this, I'm all ears :)14:04
croelandt(well, eyes since this is IRC)14:04
rosmaitathis is just for that cross-glance-tox-functional job?14:05
croelandtyes14:05
croelandtand just on 2024.214:05
croelandtso, moving on, we can come back to this after the meeting14:08
croelandt#topic One easy patch per core dev 14:08
croelandt#link https://review.opendev.org/c/openstack/glance/+/93276814:08
croelandtI could probably Ninja approve this doc/sample patch, but always good to have more eyes on a patch :)14:09
croelandt#topic Glance specs14:09
croelandt#link https://review.opendev.org/q/project:openstack/glance-specs14:09
croelandtwe can always review these before next week :)14:09
croelandt#topic Open Discussion14:09
croelandtMostly about PTG this week14:09
croelandt#link https://etherpad.opendev.org/p/oct2025-ptg-glance14:09
croelandt#link https://etherpad.opendev.org/p/2026.1-ptg-glance-planning14:09
croelandtThe proposed agenda. Please ping me if I missed anything, if a session should last longer, etc.14:10
croelandtAfter Friday EOD (my time) it'll be too late and you'll have to wait 6 months :D14:10
croelandtAnything else?14:10
mhenregarding image encryption, we have two sessions now correct?14:11
rosmaitasorry, i am in another meeting, will look at the functional test failure in 20 min or so14:12
mhenThursday @ 18 UTC Glance+Nova14:12
mhenand the other one (Glance+Cinder) does not have a fixed time slot yet or did I miss something?14:13
croelandtmhen: there is one big Nova/Glance/Cinder session, for an hour14:13
croelandtdid you want an hour with Nova and an hour with Cinder?14:13
croelandtand it's 16:00 UTC14:13
croelandtnot sure where you see 18UTC14:13
mhenoh yea 16 sorry, messed up time conversion14:13
mhenI'm fine with a combined session with all teams, that's even better14:14
mhenI wonder if 1h is enough then though14:14
rosmaitayeah, we ususally lose some time with everybody showing up, getting into the right meeting room, etc14:15
croelandtMore than an hour you usually lose people, I'd say14:15
croelandtWe're going to spend between 10 and 20 hours in technical meetings this week14:15
croelandtit's draining14:15
rosmaitayeah, but you can say, we can have two 1 hour sessions, or 1 combined 90 min session14:15
croelandtI'd say if you need more than one hour to explain your topic, it's already not great14:16
croelandtIs there no way to discuss the main ideas in one hour14:16
croelandtand refine the details during the cycle, with specs & patches?14:16
mhenJosephine (Luzi) did already meet a bunch of people and had some early discussions last week so maybe we have enough of a headstart going into the discussion14:17
rosmaitaok, stick with an hour then14:17
croelandtI'm really worried spending more than an hour is going to lead to people being fed up with the topic and not fully engaged mentally14:18
croelandtbut mhen knows better than me 14:18
croelandtmhen: if more than an hour is needed, we need to make this happen *now*14:18
mhenthe thing is: I actually don't know - up until recently this has been mostly between us, Glance and Cinder for the longest time14:19
croelandtI know :) It is tough to chair a session, especially on a super technical topic like yours14:20
mhennow Nova is joining the discussions more actively again (which is good)14:20
mhenbut I don't know how many things will be brought up or how controversial the discussion will get now that we have 3 projects discussing it and maybe 3 different perspectives on it14:21
croelandtIf this is such a huge topic, one of the key points to discuss is going to be "how do we keep in touch during the cycle?" and maybe that means we need someone from Nova and Cinder to join this meeting every week (not necessarily rosmaita and dansmith )14:21
croelandtotherwise we'll discuss it once every 6 months at PTG and it might go stale14:21
mhenwe did a weekly popup team IRC meeting for a year or two in the past but it slowly lost participants14:22
mhenmaybe we can revive that14:22
mhenI will try to discuss this in the session briefly14:22
croelandtI think you can also prepare a detailed description of the topic to make it easier to follwo along and so everyone has a rough idea of how we're doing timewise14:24
croelandtAnyway, I think we can wrap up, unless there are any other topics?14:24
rosmaitarelease patches14:24
croelandtoh yeah14:25
croelandtpatcheS, plural?14:25
croelandtI see https://review.opendev.org/c/openstack/releases/+/96422714:25
croelandtwhich is soft-blocked by the failure discussed above14:25
rosmaitaso, this situation is that one needs to be revised (update the hash)14:26
croelandtyes14:26
croelandtI would ideally manage to merge the stable/2024.2 and stable/2024.1 patch, and then change the hash14:27
rosmaitathe more recent releases only need to include stuff that is being included in Caracal but hasn't been released yet in the newer branches14:27
croelandtI don't follow14:27
rosmaitawe don't have to release from HEAD of the more recent stable branches14:28
rosmaitathough for 2024.2 it looks like we do14:28
rosmaitaok, let me back up14:29
rosmaitathis looks ready: https://review.opendev.org/c/openstack/releases/+/96424014:30
rosmaitaand so does this one: https://review.opendev.org/c/openstack/releases/+/96424114:30
rosmaitaso even if the 2024.2 gate is broken right now, we can still release14:30
croelandtyes14:31
croelandtmy concern is more Caracal14:31
croelandthttps://review.opendev.org/c/openstack/releases/+/96422714:31
croelandtthis is the *last* release for Caracal, right?14:31
rosmaitaright14:32
croelandtso if we don't merge the patches in the 2024.1 branch, they will never be part of a release14:32
rosmaitawhat else did you want to get into it?14:32
croelandtthe patch above that failed to merge for 2024.214:32
croelandthttps://review.opendev.org/q/I436d56d0489b081822ed55108e10d43cc38512bb14:32
abhishekk_@dansmith: o/ i have pushed new changes for s3 as per suggestions, please have a look when you get time14:36
dansmithabhishekk_: I know,  I have it open...just keep getting interrupted :/14:37
croelandtrosmaita: ok let's end the meeting and look at the failure "offline", ok?14:37
rosmaitaok14:38
abhishekk_no problem, thanks for the inputs, it was better and covering all the scenarios14:38
croelandt#endmeeting14:38
opendevmeetMeeting ended Thu Oct 23 14:38:38 2025 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:38
opendevmeetMinutes:        https://meetings.opendev.org/meetings/glance/2025/glance.2025-10-23-14.01.html14:38
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/glance/2025/glance.2025-10-23-14.01.txt14:38
opendevmeetLog:            https://meetings.opendev.org/meetings/glance/2025/glance.2025-10-23-14.01.log.html14:38
mhencroelandt: regarding "prepare a detailed description of the topic" - is there a good way to reach most of the session participants beforehand?14:39
mhenI could prepare a short etherpad with a topic summary, the current state and things to discuss like a lightweight agenda14:39
croelandtmhen: update your topic in https://etherpad.opendev.org/p/2026.1-ptg-glance-planning#L128 and I'll reuse that in the agenda14:41
abhishekk_mhen: you can reach via mail to openstack mailing list, keep [ptg] in subject line14:41
mhencroelandt: should I cram it directly in there or link to a dedicated etherpad?14:42
croelandtcram it in there, please14:42
mhengotcha14:42
croelandtrosmaita: I approved your releases patches15:00
rosmaitanot sure that's a good idea15:01
croelandtany idea about the test failure? Or do you someone who could help with that?15:01
croelandtrosmaita: DAMN IT15:01
croelandtwhy is that? :)15:01
rosmaitai'm not sure we caught all the changes ... trying to fiugre out the cross-functional failure now15:02
rosmaitaactually, it will only be an issue if we get this change merged into caracal, i think15:03
croelandtI, once again, don't follow15:04
rosmaitawell, if we get it into the final caracal release, it should also be in a dalmatian release, but the dalmatian release patch is pointing to a hash that doesn't include that change15:06
croelandtoh yeah15:07
rosmaitanot sure what the epoxy situation is15:07
croelandtok15:07
croelandtlet's figure out the failure then15:07
croelandtand think about releases later15:07
rosmaitayep15:07
rosmaitaso un-approve those patches, please15:08
croelandtyep, done15:11
rosmaitacroelandt: i can't find any difference between that branch and the others with respect to that job ... last recheck was a week ago, i just issued another15:38
rosmaitaand it failed15:48
dansmithabhishekk_: I think there's a problem, I commented, let me know if I'm wrong15:58
dansmithabhishekk_: much better structure of making the two paths the same though I think15:59
croelandtrosmaita: yeah maybe I should ask for help on openstack-infra?16:37
rosmaitai have asked in there for suggestions16:38
rosmaitai need to go grab lunch if you could keep an eye on the conversation16:40
rosmaita(and walk the dog, so i will be afk)16:41
croelandtoh rihgt16:41
croelandtwell I'm gonna need torun myself, but I'll stay for a bit :)16:46
croelandtrosmaita: let's rename it to glancestore17:41
rosmaitaa bit late for that, but yeah!17:42
croelandtok I'm going to angrily run now17:42
rosmaitacroelandt: ok, soon as this is fixed, i will backport that patch to Caracal and revise all the release patches17:43
croelandtso this patch is already up for Caracal17:43
croelandthttps://review.opendev.org/q/I436d56d0489b081822ed55108e10d43cc38512bb17:44
croelandtMy issue is that the deadline for Caracal is tomorrow17:44
rosmaitawell, you know how deadlines are17:51
opendevreviewBrian Rosmaita proposed openstack/glance_store stable/2024.2: Support multipart upload when image_size == 0  https://review.opendev.org/c/openstack/glance_store/+/96187018:48
opendevreviewAbhishek Kekane proposed openstack/glance master: Lazy update S3 URL on credential rotation  https://review.opendev.org/c/openstack/glance/+/96387521:21
abhishekktonyb: i think both are still relevant however we have workaround if we don't want to pin the version, also we are going to discuss this in PTG, https://review.opendev.org/c/openstack/glance-specs/+/96224821:23
tonybOkay, so it's still relevant, and the work long term-fix is probably a little way off. So it may be worth adding a cap.21:38
tonybWhere do the environment variables need to be set?  Could you add os.environ[xxx]='WHEN_REQUIRED' to the early start-up code for glance* and python-glanceclient?21:41
abhishekkthat needs to be checked, I think croelandt can answer this better21:44
tonybOkay I'll wait for them to come back online ... I'm guessing from the name they're in the EU21:46
abhishekkyeah21:48
* abhishekk going offline21:49

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