fungi | openinfra foundation board of directors meeting starting in about 6 minutes: https://board.openinfra.dev/meetings/2024-10-08 | 14:53 |
---|---|---|
fungi | "The board will receive an application presentation for a new platinum membership to the OpenInfra Foundation." | 15:00 |
fungi | cloudnull is in a suit. sort of disturbing | 15:01 |
gouthamr | :D sorta hinting on who that might be? | 15:02 |
fungi | no idea | 15:02 |
gouthamr | zer u go | 15:09 |
elodilles | hi TC, a quick double check: 2023.1 Antelope (SLURP) should transition to Unmaintained at the end of October ( https://releases.openstack.org/ ) can i start the communication towards the community about this? I'm asking because on last week's nova team meeting bauzas started thinking about removing the oldest SLURP release (antelope) will left us only with one active SLURP release: 2024.1 Caracal | 15:24 |
bauzas | elodilles: that was an open thought to be clear :) | 15:33 |
bauzas | I need to look at the SLURP resolutions :) | 15:33 |
bauzas | https://governance.openstack.org/tc/resolutions/20220210-release-cadence-adjustment.html#example-sequence is a start | 15:34 |
gouthamr | elodilles: o/ we discussed this at a recent TC meeting: https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/X3CPSOPLXPJ5VIHFPKB4SAQNRT4NXBF2/ | 15:34 |
gouthamr | elodilles: there were no objections from the TC to begin this transition.. | 15:35 |
bauzas | the resolution is also clear, I can read it | 15:36 |
bauzas | "Support: We will expect to support both the most recent “SLURP” release as well as the one prior." | 15:36 |
bauzas | there is nothing about supporting three SLURP releases | 15:36 |
bauzas | so I just had a thought but as dansmith said, this is not really something we can assess | 15:37 |
bauzas | so unless we want to discuss that, I'm fine with moving Antelope to EM when this is planned | 15:37 |
elodilles | bauzas: so you say too, that it's OK to start the transition process of 2023.1 Antelope to Unmaintained? | 15:38 |
bauzas | yes | 15:38 |
elodilles | ACK :) | 15:38 |
elodilles | then i will start the process in the coming days | 15:39 |
elodilles | thanks gouthamr and bauzas for confirming this o/ | 15:39 |
gouthamr | thanks for checking elodilles | 15:42 |
gouthamr | > so unless we want to discuss that, I'm fine with moving Antelope to EM when this is planned | 15:44 |
gouthamr | bauzas: maybe we can bring up that language during the Open Discussion at the meeting today? The language is probably vague and we can clarify | 15:44 |
bauzas | what language are you talking about ? the specifics terms of SLURP ? | 15:45 |
bauzas | and EM ? | 15:45 |
bauzas | the acronyms, I mean | 15:45 |
gouthamr | no no; that we'd support the past two SLURP releases | 15:45 |
bauzas | ah | 15:45 |
bauzas | sure, if you want to add it into the agenda | 15:45 |
elodilles | i might not be around but will check the meeting logs o:) thanks again o/ | 15:47 |
gouthamr | ++ | 15:48 |
opendevreview | Goutham Pacha Ravi proposed openstack/governance master: Fix Typo in Release Cadence Adjustment resolution https://review.opendev.org/c/openstack/governance/+/931814 | 15:50 |
gmann | elodilles: sounds good to me and it is as per timeline | 15:53 |
elodilles | gmann: ACK, thanks o/ | 15:55 |
gouthamr | tc-members: gentle reminder that we're meeting here in ~40 minutes | 17:21 |
gouthamr | #startmeeting tc | 18:00 |
opendevmeet | Meeting started Tue Oct 8 18:00:13 2024 UTC and is due to finish in 60 minutes. The chair is gouthamr. Information about MeetBot at http://wiki.debian.org/MeetBot. | 18:00 |
opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 18:00 |
opendevmeet | The meeting name has been set to 'tc' | 18:00 |
cardoe | \o | 18:00 |
gouthamr | Welcome to the weekly meeting of the OpenStack Technical Committee. A reminder that this meeting is held under the OpenInfra Code of Conduct available at https://openinfra.dev/legal/code-of-conduct. | 18:00 |
gouthamr | Today's meeting agenda can be found at https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee | 18:00 |
gouthamr | #topic Roll Call | 18:00 |
gmann | o/ | 18:00 |
bauzas | \o | 18:00 |
slaweq | o/ | 18:00 |
noonedeadpunk | o/ | 18:00 |
cardoe | \o | 18:00 |
gouthamr | noted absence: f r i c k l e r | 18:01 |
gouthamr | courtesy ping: spotz[m] gtema | 18:02 |
spotz[m] | o/ | 18:02 |
spotz[m] | ha! | 18:03 |
gtema | o/ | 18:03 |
gouthamr | awesome; that's more than the quorum.. thanks for joining; lets get started | 18:03 |
gouthamr | #topic Last Week's AIs | 18:04 |
gouthamr | 1) Respond to the Watcher mail thread and ask for an update this week (gouthamr) | 18:04 |
gouthamr | so i did this; but not early enough :/ | 18:04 |
gouthamr | #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/3DRYZFDPVCZ45TOULOZ4R7K6BUOIHLU2/ ([tc][watcher] No leaders for project team, heading to retirement) | 18:05 |
gouthamr | sean-k-mooney isn't here; but will poke him where he is | 18:05 |
gouthamr | so lets get some status between this meeting and the next one there | 18:06 |
gouthamr | 2) Review the patch for marking Kuryr-related projects inactive this week (Patch: https://review.opendev.org/c/openstack/governance/+/929698) (everyone) | 18:06 |
gouthamr | we have sufficient votes here; so i'll press the workflow button here | 18:07 |
gouthamr | 3) Share the PTG planning etherpad on the mailing list and ask interested participants to vote on topics, indicating their time zones (gouthamr) | 18:07 |
gouthamr | #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/message/YDG3VCGHGRMZ7KBPW4JWJDXKO4OUZIE7/ ([all][tc][ptl][ptg] TC Discussions at the PTG - Sign Up and Suggest Topics) | 18:07 |
gouthamr | ^ not a lot of sign ups there; but, we have timeslots for our sessions now | 18:08 |
gouthamr | #link https://ptg.opendev.org/ptg.html (PTG Schedule) | 18:08 |
gouthamr | the TC meetings are scheduled 21st Oct: 1400 UTC-1700 UTC and 25th Oct: 1500 UTC to 1700 UTC | 18:09 |
gouthamr | lets dive into this in a little bit | 18:10 |
bauzas | noted, I'll ask the nova community for leading their sessions by someone else | 18:10 |
gmann | do we have onyl leaders interaction session on Monday or more slots? | 18:10 |
gmann | usually Thursday and friday works fine for TC related discusison | 18:11 |
opendevreview | Merged openstack/governance master: Mark kuryr-kubernetes and kuryr-tempest-plugin Inactive https://review.opendev.org/c/openstack/governance/+/929698 | 18:11 |
gouthamr | ^ please hold that thought gmann | 18:12 |
gmann | especially friday 17 UTC or 18 UTC slot where more community members can join after finishing their project discussions | 18:12 |
gmann | sure | 18:12 |
gouthamr | that's all the AIs I was tracking; was there anything else being worked on during this past week | 18:12 |
noonedeadpunk | fwiw, I'm traveling Thursday afternoon, but back at evening | 18:12 |
gouthamr | there was one about OSC requiring a new release and a upper-constraint bump | 18:12 |
gmann | there was on AI on me I think for migration to ubuntu noble goal | 18:13 |
gmann | I have created a etherpad and starting working on that | 18:13 |
gmann | #link https://etherpad.opendev.org/p/migrate-to-noble | 18:13 |
gouthamr | ack thank you | 18:13 |
gmann | should be able to prepare the base changes to test by this week and send it on ML | 18:14 |
gouthamr | thanks for adding that to the tracker too! | 18:14 |
gouthamr | ++ | 18:14 |
gouthamr | #link https://bugs.launchpad.net/python-openstackclient/+bug/2080600 (openstack user create fails without --domain option) | 18:15 |
gouthamr | #link https://review.opendev.org/c/openstack/python-openstackclient/+/930412 ( identity: Don't pass unset options when creating user - stable/2024.2) | 18:15 |
gouthamr | #link https://review.opendev.org/c/openstack/requirements/+/931811 (update constraint for python-openstackclient to new release 7.1.3) | 18:16 |
bauzas | ush... | 18:16 |
gouthamr | ^ that's on master though.. | 18:16 |
gouthamr | sry | 18:16 |
gouthamr | scratch that | 18:16 |
gouthamr | 2024.2 | 18:16 |
gtema | release change was merged few hours bach | 18:17 |
gmann | yes, master one need release and u-c update but meanwhile fix can be done by passing the domain-id | 18:17 |
gmann | ok | 18:17 |
gmann | I did for greande master | 18:17 |
gmann | #link https://review.opendev.org/c/openstack/grenade/+/931057/3 | 18:17 |
gmann | but it will be good to have that in u-c also | 18:17 |
gouthamr | great; we have crossed the i's and dotted the t's | 18:18 |
bauzas | so for grenade, what's that mean ? | 18:18 |
gmann | it create users and was not passing domain id | 18:18 |
* bauzas opens the grenade change | 18:18 | |
gouthamr | in the scripts, if you didn't specify a domain ID, you'd hit that bug | 18:18 |
gmann | gtema: one question, why it was not caught when change introduced the bug ? | 18:19 |
gmann | are we lacking some testing on requirements when u-c was bumped? | 18:19 |
gmann | I think this is change introduced it? | 18:20 |
gmann | #linl https://review.opendev.org/c/openstack/python-openstackclient/+/909030 | 18:20 |
gmann | #link https://review.opendev.org/c/openstack/python-openstackclient/+/909030 | 18:20 |
bauzas | yeah, that's a good question | 18:20 |
gtema | because we are not testing all the weird and wrong usages | 18:20 |
bauzas | I wonder why we didn't found it | 18:20 |
bauzas | we already had a problem with OSC before RC1 | 18:20 |
gmann | devstack and grenade job could have caught it easily | 18:20 |
gmann | and requirement gate run tempest-full job | 18:21 |
gmann | not sure why it passed there? | 18:21 |
gmann | sorry tempest-full would not catch it but grenade job and swift job will | 18:21 |
gmann | swift-dsvm-functional was failing on devstack | 18:22 |
gmann | I think we should add grenade job also in requirement gate to test more coverage | 18:22 |
bauzas | but aren't we testing a OSC change by a job that calls the APIs ? | 18:22 |
gmann | bauzas: that would not catch it until it is in u-c | 18:23 |
gmann | unless we are testing it with master? | 18:23 |
bauzas | because tempest doesn't use OSC, right? | 18:23 |
gmann | yeah | 18:23 |
bauzas | ... | 18:23 |
gmann | I think swift-dsvm-functional and adding grenade job should catch osc related things there | 18:24 |
gmann | swift one use the osc | 18:24 |
bauzas | so when we merge a OSC change, we only actually test it once we have a new release.. ouch. | 18:24 |
gmann | I can propose change to increase the coverage in requirements gate | 18:25 |
bauzas | probably but I would prefer to check that by a OSC job | 18:25 |
gouthamr | +1 on adding a grenade job; /me is thinking if we will end up blocking something inadvertently if that job was voting | 18:25 |
bauzas | hopefully by the check pipeline, but a periodic could work | 18:26 |
bauzas | because once OSC releases a version, someone can use it | 18:26 |
gmann | bauzas: osc job might not be testing all these APIs but the osc tests etc can be changed in the same change when backward incompatible change is introduced | 18:27 |
gtema | but you can't also test everything in osc gate | 18:27 |
gmann | yeah | 18:27 |
bauzas | everything for sure | 18:27 |
bauzas | but nova, neutron, cinder ? | 18:27 |
gmann | gouthamr: how about testing upgrade with grenade job? | 18:28 |
bauzas | yeah that'd find it | 18:28 |
gmann | bauzas: ++ and these few projects like we do in requirements and oslo libs | 18:28 |
gmann | their functional jobs should run quickly an catch the things in advance | 18:28 |
bauzas | at least I'd prefer to have a running OSC job before releasing a version | 18:29 |
gmann | for example olso.policy test nova, neutron tox and functional job and we get to know if any breaking things happening | 18:29 |
bauzas | that's why I said we could run periodics if the OSC commnunity can't or doesn't want to have check votes | 18:30 |
bauzas | but before releasing, they would check that periodic | 18:30 |
gmann | a few checks job should not harm even | 18:30 |
bauzas | sure the more the better | 18:30 |
gmann | I can propose changes and we can see how it looks like | 18:30 |
gouthamr | ++ thanks gmann | 18:31 |
gouthamr | alright lets wrap up the AIs with that note; and move on to regular programming | 18:31 |
gouthamr | #topic TC PTG | 18:31 |
bauzas | gmann++ agreed | 18:31 |
gouthamr | noonedeadpunk: noted your absence on Thursday.. but, i did mean to have TC sessions only on Mon, Fri - not an executive decision.. was just following a pattern from the past few PTGs | 18:32 |
noonedeadpunk | ++ | 18:32 |
gouthamr | on Monday; however, i reserved more time because i think it'd be helpful to have project contributors join us on the goal topics on the etherpad | 18:32 |
bauzas | yup and, | 18:33 |
gmann | gouthamr: note we do have other support team sessions on Modnay and people may want to join there | 18:33 |
gmann | QA usually book on Monday and try to finish it before members join other project related discussion on other days | 18:34 |
bauzas | for example if we discuss about the translation topic in Monday, projects could also engage that within their own PTG times after | 18:34 |
gouthamr | gmann: i didn't see any yet.. probably still planning? | 18:34 |
spotz[m] | The problem is no one has signed up on the schedule yet | 18:34 |
gouthamr | ^ YES | 18:34 |
gmann | gouthamr: yes | 18:34 |
gmann | I will check with martin on that | 18:34 |
spotz[m] | I will miss part of Monday but will pop in and out | 18:34 |
bauzas | I like the idea of a TC starting cross-project topics at the beginning of the PTG and eventually concluding at the end | 18:34 |
gmann | but why we do not book Friday evening slot for TC? | 18:34 |
gouthamr | sigh; okay.. we'll try to be dynamic about this.. maybe we don't need 3 hours on Monday.. we could do with 2 | 18:35 |
gouthamr | gmann: it is? | 18:35 |
gmann | that was most attended slots in past | 18:35 |
gouthamr | ohh, you mean the very last slot | 18:35 |
gmann | yeah, maybe 2 hrs or 3 hrs the last slots on friday | 18:35 |
bauzas | I'd like to see some topics engaged on Monday, 2 hours can be a stretch goal but I'd prefer to keep the 3rd hour booked | 18:36 |
gmann | when community members finish project related things and can join TC one | 18:36 |
gmann | in that case, maybe we can shift Thursday one to Friday? | 18:36 |
gouthamr | hmmm, 1700 UTC; past beer-o-clock in EU and APAC, and nearly next day far east | 18:37 |
fungi | some of it stems from in-person ptgs where lots of people were already leaving to catch their flights on friday afternoons, but still possible some people are dropping out to start other weekend activities (especially in europe, and in asia it's already saturday at that point). personally, i have some outside obligations on that friday and won't be arounbd | 18:37 |
gmann | bauzas: sure | 18:37 |
bauzas | the Friday slots will impact my personal activities, for sure, but that's only once per semester | 18:37 |
gmann | you mean friday last slots are not recommened to book? | 18:37 |
spotz[m] | If we were in person sure, but you're more likely to take a Friday off then most other days | 18:38 |
gmann | for virtual i mean, there is no in person PTG now | 18:38 |
spotz[m] | And we run into it's evening for EMEA and Saturday for APAC as mentioned | 18:38 |
bauzas | trust me, virtual PTGs are very in-person for me :) | 18:39 |
bauzas | the only difference is that I'm alone in my room | 18:39 |
gmann | :) | 18:39 |
bauzas | didn't we had those usual timeslots in the past ? | 18:40 |
bauzas | I'm fine with keeping me, I was prepared :) | 18:40 |
bauzas | keeping them* | 18:40 |
slaweq | yeah, we had usually Monday for TC & community leaders session and then Thursday and Friday for TC discussion | 18:41 |
bauzas | but we may require some asia friendly timeslot in order to balance | 18:41 |
gmann | yeah | 18:41 |
slaweq | it was up to around 18 or 19 utc IIRC | 18:41 |
bauzas | if you want MHO, | 18:41 |
bauzas | booking a slot doesn't mean we need to use it fo the whole time | 18:41 |
fungi | (noting it's almost 19 utc now) | 18:41 |
gouthamr | true ^ but, if we moved this meeting to friday, how many of us would be happy to join in? | 18:42 |
gmann | asia friendly slots is good idea but even 13 UTC is not asia friendly right? | 18:42 |
bauzas | I'm used to book 16 hours every cycle which I usually take, but we had occurences in the past of finishing earlier, and this was fine | 18:42 |
gmann | gouthamr: weekly meetings vs PTG are different things. | 18:42 |
slaweq | gouthamr I wouldn't be happy but I would join :) | 18:42 |
bauzas | gouthamr: a weekly meeting is different from a once-semesterly (is it the right term) meeting ? | 18:42 |
bauzas | gmann: that. | 18:43 |
bauzas | I'm OK with running late on a Friday evening once, I'm not okay with throwing *all* my Friday evening activities the whole 6 months | 18:43 |
gmann | we do not need to book 18 or 19 UTC on friday even 15-17 UTC 2 hrs should be good | 18:44 |
slaweq | bauzas yes, the same for me, once in 6 months I can definitely do it | 18:44 |
spotz[m] | Yeah but a weekly on Friday isn't the most productive as at any given time 1-3 people would be off | 18:44 |
gmann | even I like the 18-19 UTC one which has been productive in past but anyways | 18:44 |
bauzas | did we got updates from ian choi and sungsoo about their topics ? | 18:44 |
slaweq | but if we want to have some time slots more Asia friendly, then I guess it will be somewhere in the middle of night for Europe | 18:45 |
bauzas | given the productive Korea summit, I just want to make sure we leave room for them | 18:45 |
gouthamr | gmann: 1300 UTC is 9pm in Beijing/Manila/Perth, 10pm in Tokyo, 12am in Sydney -- not exactly "friendly"; but 0400-0800 would be hard for EU and west asia folks | 18:45 |
gmann | gouthamr: there is no time we can accommodate all these TZ :) | 18:45 |
bauzas | if that's once *and* productive, I can manage my sleep miss | 18:45 |
bauzas | let's run a physical PTG ! | 18:46 |
slaweq | gouthamr personally I can do it also in such weird time slots if needed, it is just once in 6 months and I would feel more like on in-person PTG, with jetlag :D | 18:46 |
gmann | maybe we can open alternate slot for them but finding slot which is ok for all these TZ is not possible | 18:46 |
gmann | friday 15-17 UTC can be good one and not very late ? | 18:47 |
bauzas | (I should say a physical-in-the-same-room-term PTG | 18:47 |
bauzas | gmann: as I said, I'll just need to find someone to chair the nova sessions but I'm cool with the current proposal | 18:48 |
gouthamr | ack; from what i'm reading, i'd throw in the session times with the existing time slots, and have some wiggle room for cross project discussions on Monday outside the TC room | 18:48 |
gouthamr | we can do one final check next week; and move things around | 18:48 |
bauzas | ++ | 18:48 |
gouthamr | does that make sense? short notice, but, we're used to that | 18:48 |
gmann | ++ | 18:48 |
gouthamr | alright anything else about the PTG? | 18:50 |
spotz[m] | And maybe an email to the list of what we're planning though maybe that's why no one scheduled Monday in anticipation | 18:50 |
gmann | usually teams are late to book | 18:50 |
gouthamr | ^ yes will update my thread stating there are now time slots | 18:50 |
gouthamr | (when they're there) | 18:50 |
cardoe | I need to book something for a neutron/ironic cross session | 18:51 |
gouthamr | #topic A check on gate health | 18:52 |
gouthamr | we | 18:52 |
gouthamr | are 8 mins away from close | 18:52 |
gouthamr | and i'd like to keep a few for Open Discussion | 18:52 |
gmann | greande jobs is not yet ready to upgrade from stable/2024.2 to current master | 18:52 |
gmann | #link https://review.opendev.org/c/openstack/grenade/+/930507/8 | 18:53 |
clarkb | opendev is going to update the default ansible version to 9 in the openstack tenant (and all other tenants) today | 18:53 |
clarkb | *in the openstack zuul tenant | 18:53 |
gmann | some setuptool error is happening in multinode jobs | 18:53 |
gmann | #link https://zuul.opendev.org/t/openstack/build/517b21fc57d6480f902273e58f9032b5/log/compute1/logs/old/devstacklog.txt#5046 | 18:53 |
gmann | anyone aware of these? | 18:53 |
gmann | clarkb: fungi ^^ | 18:53 |
clarkb | gmann: looks like it is trying to do an editable install and that is failing. Did openstack stop doing editable installs elsewhere or maybe found some other workaround? | 18:54 |
fungi | that error doesn't look familiar at least | 18:54 |
fungi | but there were semi-recent (in the past 1-2 years) changes to how editable installs are handled by setuptools for pyproject standardization | 18:55 |
gmann | not sure, it only happen in multinode job and when it install devstack on stable/2024.2 (as old node in greande) | 18:55 |
fungi | the changes i'm familiar with only kick in when the package has a pyproject.toml file though | 18:55 |
clarkb | fungi: it is using build_meta stuff which I think implies pyproject.toml? | 18:56 |
gmann | ok, maybe something in nova side in stable/2024.2 | 18:56 |
clarkb | I don't see any obvious changes in devstack/inc/python explaining why it generally works but doesn't with grenade | 18:56 |
gmann | but we should have some multinode job there which were passing, it started failing when greande testing it from stable/2024.2 | 18:57 |
fungi | nova added a pyproject.toml file in commit 8464abe56ccbbf09f768df47707d2ff2c4c7b264 which appears in tags 30.0.0.0rc1 and 30.0.0 | 18:58 |
fungi | not sure if that could explain it | 18:58 |
gmann | maybe, I was checking the same. thanks | 18:58 |
gouthamr | #link https://review.opendev.org/c/openstack/nova/+/899753 | 18:59 |
gouthamr | same thing on gerrit ^ | 18:59 |
fungi | it definitely appears in the stable/2024.2 branch though | 18:59 |
gmann | anyways I will check and discuss with you after meeting or so | 18:59 |
gmann | fungi: yeah | 18:59 |
gmann | fungi: let me test that | 18:59 |
fungi | how editable installs work with setuptools definitely changes depending on whether the project has a pyproject.toml file though, yes | 19:00 |
gmann | k | 19:00 |
gouthamr | alright we're at the hour | 19:00 |
gouthamr | and we're out of time for Open Discussion | 19:01 |
gouthamr | we'll try to slot this better in future meetings.. | 19:01 |
gouthamr | but, please chat away after we end this meeting | 19:01 |
gouthamr | thank you all for attending | 19:01 |
gouthamr | #endmeeting | 19:01 |
opendevmeet | Meeting ended Tue Oct 8 19:01:56 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 19:01 |
opendevmeet | Minutes: https://meetings.opendev.org/meetings/tc/2024/tc.2024-10-08-18.00.html | 19:01 |
opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/tc/2024/tc.2024-10-08-18.00.txt | 19:01 |
opendevmeet | Log: https://meetings.opendev.org/meetings/tc/2024/tc.2024-10-08-18.00.log.html | 19:01 |
slaweq | thx, see you | 19:01 |
gtema | cy | 19:02 |
gouthamr | We have OpenInfra Days NA next week (2024-10-15) | 19:02 |
noonedeadpunk | fwiw, I keep seing quite some failures in CI since a week or so, whenever we'd need to reach the world... | 19:02 |
gouthamr | noonedeadpunk: o/ was going to check if you're around during this meeting time next week, and if you can run this meeting :) | 19:03 |
noonedeadpunk | oh, yes, for sure | 19:03 |
gmann | fungi: clarkb: testing with nova revert as depends-on https://review.opendev.org/c/openstack/grenade/+/930507 | 19:03 |
gouthamr | noonedeadpunk++ thank you; i'll recap to the list as usual and prep the agenda with the AIs before hand | 19:03 |
bauzas | thanks gouthamr | 19:04 |
fungi | noonedeadpunk: we did have an outage in ovh bhs1 earlier today for a few hours, where ipv6 routing broke for a while | 19:04 |
noonedeadpunk | ++ thanks ~! | 19:04 |
fungi | but that wouldn't fit with "a week or so" unfortunately | 19:04 |
clarkb | gmann: I would look at regular devstack runs and comprae the installation and see why/how they are different | 19:05 |
clarkb | then work backward from there | 19:05 |
gmann | clarkb: yeah, I think comparing multinode job on that change which was passing there. | 19:07 |
gmann | will do | 19:07 |
fungi | opendev just merged the change to update the default ansible version in zuul too | 19:16 |
clarkb | it might take a minute to deploy (since that doesn't happen in the internal configs we have to write to disk and tell zuul to reload) | 19:18 |
gmann | ++ | 19:26 |
fungi | 19:26:14 utc: Build succeeded (deploy pipeline). | 19:28 |
clarkb | gmann: compare success https://zuul.opendev.org/t/openstack/build/f85acb9cc88e44d9908c82400738f614/log/job-output.txt#10777 to failure https://zuul.opendev.org/t/openstack/build/517b21fc57d6480f902273e58f9032b5/log/compute1/logs/old/devstacklog.txt#5023 and one thing I notice is that we're using a virtualenv for regular devstack and using system python3.10 in the fail case | 19:52 |
clarkb | I wonder if the pip install -e needs to do extra magic in the system install case because it is a system install and keeping track of editable stuff in that context is different than in a virutalenv? | 19:53 |
clarkb | generally I think grenade probably should move to a venv anyway so may be worth pulling on that thread to see if it helps | 19:54 |
clarkb | and if not the effort isn't wasted as that is an end goal anyway | 19:54 |
clarkb | in particular the virutalenv may update setuptools but system python wouldn't. That could explain why the method is not found (system setuptools is too old for the pip expectation?) | 19:55 |
gmann | clarkb: I see, I thought grenade also enabled the venv when devstack did it by default using the GLOBAL_VENV and I see we disabled it in grenade side | 20:03 |
gmann | updated the change, let's see thanks clarkb for help | 20:07 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!