| melwitt | gmann: IIUC this change made it so it's not possible to run the ceph jobs on branches older than train (devstack-plugin-ceph branching starts at stable/train) https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/786311 was that intentional to remove the job runs from branches older than train? | 00:19 |
|---|---|---|
| melwitt | *the devstack-plugin-ceph-tempest job runs | 00:19 |
| gmann | melwitt: checking | 00:21 |
| melwitt | I note we do have ceph enabled for the live migration and grenade multinode jobs as well, so we do have some ceph coverage even without devstack-plugin-ceph-tempest | 00:22 |
| gmann | melwitt: yeah, when we discussed about branching it in QA, we decided to do until stable/train as other older stable branch are in EM state. and as it is branched, we need to run other stable branch jobs in other stable gate | 00:23 |
| gmann | for branchless we needed stable jobs on master gate | 00:23 |
| melwitt | gmann: ok cool. I ask because I'm updating the ceph grafana page and wondering if I should remove the graphs for older than train | 00:23 |
| gmann | melwitt: currently situation is, <stable/train gate is tested with master devstack-plugin-ceph-tempest which i am not sure how long it will break | 00:24 |
| gmann | *it will work | 00:24 |
| melwitt | based on this info, answer is yes | 00:24 |
| melwitt | ack | 00:24 |
| gmann | if they are broken then maintainer for EM stable brnach can branch devstack-plugin-ceph-tempest for that older branch | 00:25 |
| gmann | yeah, for now we can cleanup grafana page | 00:25 |
| gmann | if anyone cut the older stable then it can be added back | 00:25 |
| melwitt | ok cool. yeah I just wanted to make sure it was an intended thing before I just delete the graphs heh | 00:25 |
| gmann | +1 | 00:26 |
| melwitt | gmann: actually sorry, I see that these jobs are still in here on the master branch https://opendev.org/openstack/devstack-plugin-ceph/src/branch/master/.zuul.yaml#L35-L64 but I don't understand why they don't run on stable/stein, for example https://github.com/openstack/nova/blob/stable/stein/.zuul.yaml#L252 do you know why? just wanted to know how this works if you know | 00:34 |
| gmann | melwitt: ohk, let me check. we have branch variant there so should run | 00:36 |
| melwitt | gmann: I just need to know the answer to this mystery 😆 | 00:37 |
| gmann | melwitt: zuul job definition matching is mystery many times :) especially with branch variant matching | 00:39 |
| melwitt | haha yes | 00:44 |
| melwitt | I was in the process of writing the commit message "removing these graphs because they were removed in change X" and looked and saw they are still there in the repo 😝 so I'm back to being confused | 00:44 |
| melwitt | *because their jobs were removed | 00:45 |
| gmann | I remember now we kept these stable/stein and older branch job definition for that purpose only - project stable/stein or older can keep running these from here | 00:47 |
| gmann | https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/786307/1/.zuul.yaml#65 | 00:48 |
| gmann | but seems that is not working. | 00:48 |
| gmann | checking if master devstack-plugin-ceph-tempest is checkout out there or not | 00:48 |
| melwitt | ack yeah | 00:54 |
| gmann | melwitt: I think what's happening here as devstack-plugin-ceph is not branchless anymore so stable/stein gate does not checkout the master version and so does it does not find the job definition so does not run it | 01:15 |
| gmann | stable/stein ran these jobs only until we made devstack-plugin-ceph branched - https://zuul.openstack.org/builds?job_name=devstack-plugin-ceph-tempest&branch=stable%2Fstein | 01:16 |
| melwitt | gmann: what happens if it wants to checkout stable/stein but there is no stable/stein branch? or are you saying that if no matching branch exists, it doesn't do anything? | 01:17 |
| gmann | I see two solution here 1. cut the stable/stein branch for devstack-plugin-ceph 2. define the job in project side like nova which can checkout master devstack-plugin-ceph | 01:17 |
| gmann | melwitt: yes | 01:17 |
| gmann | 2nd one we need to try if that work | 01:17 |
| melwitt | ok, thanks. good to know | 01:17 |
| gmann | but first question is if we want to run the ceph testing on EM stable/stein ? | 01:18 |
| gmann | from nova perspective. | 01:18 |
| gmann | if yes then I can cut the stable/stein in devstack-plugin-ceph | 01:18 |
| sean-k-mooney[m] | you an specify a branch override in the job definition with the branch to checkout for depends projects | 01:18 |
| melwitt | yeah. that's what I'm not sure about. I'm guessing probably not | 01:18 |
| gmann | ok | 01:18 |
| gmann | in QA we were like "if someone ask EM testing coverage then we cut the branch stable/stein or odler otherwise leave" | 01:19 |
| sean-k-mooney[m] | you should not need to cut a stable/stein branch | 01:20 |
| gmann | sean-k-mooney[m]: we do that on but on master branch of branched devstack-plugin-ceph | 01:20 |
| sean-k-mooney[m] | the stable/stein job that need a compatible version can specify an override to use | 01:20 |
| gmann | which is not working | 01:20 |
| melwitt | yeah, makes sense. mostly I wanted to know why it wasn't running bc from what I could tell it should be | 01:20 |
| gmann | sean-k-mooney[m]: yes you mean 2nd option? if we define the job in nova stable/stein then it can work | 01:20 |
| melwitt | CI magic | 01:21 |
| gmann | melwitt: k. ping me in QA if you think this coverage requirement or asked by someone then we can plan | 01:21 |
| sean-k-mooney[m] | it does not have to be in nova stable stein technically it can live in out repos | 01:21 |
| gmann | sean-k-mooney[m]: out in any other project but in stable/stein code | 01:22 |
| gmann | devstack-plugin-ceph does not have stable/stein (only until stable/train) and that is why it is complicated here | 01:22 |
| sean-k-mooney[m] | i don’t think it nessisarly need to be in stable/stein code | 01:22 |
| gmann | sean-k-mooney[m]: here we have currently which does not work https://opendev.org/openstack/devstack-plugin-ceph/src/branch/master/.zuul.yaml#L35-L64 | 01:23 |
| melwitt | gmann: ack. I'll ask lyarwood next week for sanity check | 01:23 |
| gmann | +1 | 01:23 |
| sean-k-mooney[m] | right but you can do something like this https://github.com/SeanMooney/ci-sean-mooney/blob/main/zuul.d/jobs.yaml#L30-L42 | 01:24 |
| sean-k-mooney[m] | that is not using override checkout | 01:25 |
| melwitt | oh neat | 01:25 |
| melwitt | I might try it in DNM patch just to see | 01:25 |
| gmann | we can try if that work | 01:25 |
| sean-k-mooney[m] | in this case i was using it to translate between my external ci branch which was main and master | 01:26 |
| gmann | override_checkout can checkout master for stable/stein variant of job but if that is identified in stable/stein gate or not is something we need to test | 01:26 |
| melwitt | sorry I didn't mean to make this sound like a problem, I just wanted to understand what's going on | 01:26 |
| gmann | sean-k-mooney[m]: if only master then it work like before it was working in case of devstack-plugin-ceph being branchless | 01:26 |
| gmann | melwitt: sean-k-mooney[m] and other problem is that is we will test stable/stein with master devstack-plugin-ceph which can be broken anytime. | 01:27 |
| sean-k-mooney[m] | so i think overdid checkout can be any valid git checkout | 01:27 |
| sean-k-mooney[m] | like a git commit sha tag or branch | 01:27 |
| gmann | but if something broken then we have to cut the branch even tag checkout or hash | 01:28 |
| melwitt | gmann: that's a good point | 01:28 |
| sean-k-mooney[m] | so we can have the stable/stein version of the job checkout a version of devstack-plugin-ceph that works with that code | 01:28 |
| gmann | I will say if we want to test stable/stein then we should cut the branch | 01:28 |
| sean-k-mooney[m] | if you intend to release it as a branched project then yes | 01:29 |
| sean-k-mooney[m] | i’m not really going to argue for or against testing it be i belive e/// has a product based on stein so gibi is proably somewhat interested in it | 01:29 |
| gmann | we left that until it is needed and with someone volunteering for maintaining it :) | 01:29 |
| gmann | that is why we did cut until stable/train only | 01:30 |
| sean-k-mooney[m] | i don’t think that makes sense | 01:30 |
| sean-k-mooney[m] | either you should alswasy branch or at least tag or never | 01:31 |
| gmann | it is converted from branchless to branched recently so we left EM stable branches until if anyone need and help to maintain that | 01:31 |
| gmann | for regular branched repo yes we should cut always but this is different case | 01:32 |
| sean-k-mooney[m] | ah ok | 01:32 |
| sean-k-mooney[m] | will the stable train branch work with stable/stein | 01:32 |
| melwitt | yeah this was retro branched | 01:32 |
| gmann | currently yes but if stable/train has some change then it can break | 01:33 |
| sean-k-mooney[m] | only if we don’t gate on stone passing with it | 01:34 |
| gmann | there is no extra cost of cutting stable/stein also but we just need to maintain it when it break | 01:34 |
| sean-k-mooney[m] | sure althoughif its EM we wont be landing patches there really | 01:34 |
| gmann | we do backport in EM branch also in same way at least for stable/stein | 01:35 |
| gmann | anyways we can check if gibi lyarwood interested to test it then we can cut branch | 01:36 |
| sean-k-mooney[m] | right but not without testing right and only if the bug was reported against that branch | 01:36 |
| gmann | yeah | 01:37 |
| sean-k-mooney[m] | something like a devstack plugin should have very few backports as really they are only required if the distribution breaks something | 01:37 |
| sean-k-mooney[m] | after all its primary reason for exsiting is to enable ci and that way its used in those ci jobs should not change much after the release is made | 01:38 |
| opendevreview | melanie witt proposed openstack/nova stable/stein: DNM testing https://review.opendev.org/c/openstack/nova/+/799157 | 02:54 |
| opendevreview | Yongli He proposed openstack/nova master: Smartnic support - cyborg drive https://review.opendev.org/c/openstack/nova/+/771362 | 03:03 |
| opendevreview | Yongli He proposed openstack/nova master: smartnic support - new vnic type https://review.opendev.org/c/openstack/nova/+/771363 | 03:03 |
| opendevreview | Yongli He proposed openstack/nova master: smartnic support - create arqs https://review.opendev.org/c/openstack/nova/+/758944 | 03:03 |
| opendevreview | Yongli He proposed openstack/nova master: smartnic support - build instance with smartnic arqs https://review.opendev.org/c/openstack/nova/+/798249 | 03:03 |
| opendevreview | Yongli He proposed openstack/nova master: smartnic support - cleanup arqs https://review.opendev.org/c/openstack/nova/+/798054 | 03:03 |
| opendevreview | Yongli He proposed openstack/nova master: smartnic support - reject server move and suspend https://review.opendev.org/c/openstack/nova/+/779913 | 03:03 |
| opendevreview | Yongli He proposed openstack/nova master: smartnic support - functional tests https://review.opendev.org/c/openstack/nova/+/780147 | 03:03 |
| yonglihe | commit "Add --task-log option to nova-manage db archive_deleted_rows" introduced new functional test case "test test_archive_task_logs ", seems failed while checking logs. | 06:20 |
| gibi | sean-k-mooney[m]: we have product on mitaka (lol), pike, (train?), victoria | 07:52 |
| gibi | but we are not using ceph afaik | 07:53 |
| sean-k-mooney | gibi: i tought ye had one on stien too but maybe i was thinking of the pike version. i rememebered you had a release that was one older then the release we needed for OSP | 09:42 |
| sean-k-mooney | i rememebre asking if you were going to backport a fix before to a release and you were going one older for your product | 09:43 |
| opendevreview | sean mooney proposed openstack/os-vif master: update os-vif ci to account for devstack default changes https://review.opendev.org/c/openstack/os-vif/+/798038 | 10:03 |
| opendevreview | sean mooney proposed openstack/os-vif master: add configurable per port bridges https://review.opendev.org/c/openstack/os-vif/+/798055 | 10:03 |
| opendevreview | sean mooney proposed openstack/os-vif master: update os-vif ci to account for devstack default changes https://review.opendev.org/c/openstack/os-vif/+/798038 | 10:06 |
| opendevreview | sean mooney proposed openstack/os-vif master: add configurable per port bridges https://review.opendev.org/c/openstack/os-vif/+/798055 | 10:06 |
| opendevreview | Lee Yarwood proposed openstack/nova master: zuul: Add CentOS 8 stream integrated compute tempest job to gate https://review.opendev.org/c/openstack/nova/+/797616 | 10:45 |
| *** bhagyashris_ is now known as bhagyashris|ruck | 12:15 | |
| lyarwood | melwitt: https://bugs.launchpad.net/nova/+bug/1934519 - when you're back online on Tuesday btw, not sure why this has started to fail 100% of the time. | 13:54 |
| ralonsoh | gibi, hi, can you join neutron channel? | 14:15 |
| ralonsoh | we are talking about https://bugs.launchpad.net/neutron/+bug/1930866 | 14:16 |
| opendevreview | Lee Yarwood proposed openstack/nova master: block_device: Use initialize APIs to refresh when reported as idempotent https://review.opendev.org/c/openstack/nova/+/720769 | 15:19 |
| opendevreview | Julia Kreger proposed openstack/nova master: [ironic] Minimize window for a resource provider to be lost https://review.opendev.org/c/openstack/nova/+/799327 | 20:03 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!