Wednesday, 2024-08-21

*** __ministry is now known as Guest102102:02
*** bauzas_ is now known as bauzas03:51
*** bauzas_ is now known as bauzas06:34
*** elodilles_ooo is now known as elodilles06:44
*** bauzas_ is now known as bauzas07:34
gboutryHello, is the mirror provided in ansible var docker_mirror only a mirror for docker.io or can it mirror other registries as well?08:28
opendevreviewdaniel.pawlik proposed openstack/ci-log-processing master: Add url subpath for Logsender to reach OpenSearch  https://review.opendev.org/c/openstack/ci-log-processing/+/92671709:31
opendevreviewdaniel.pawlik proposed openstack/ci-log-processing master: Add url subpath for Logsender to reach OpenSearch  https://review.opendev.org/c/openstack/ci-log-processing/+/92671709:31
fricklergboutry: iiuc it is not even a mirror, but only a caching proxy. you could propose a patch to add additional proxy settings if needed09:32
gboutryWhich repo contains the tasks to deploy this caching proxy? Would like to see how it's setup09:35
fricklergboutry: this should have most of the config https://opendev.org/opendev/system-config/src/branch/master/playbooks/roles/mirror/templates/mirror.vhost.j209:40
gboutryfrickler: thank you very much, I can see a quay registry mirror, but not variable like docker_mirror to expose it, are we supposed to construct this value in our own ansible tasks?09:45
gboutry(just trying to determine a pattern)09:46
fricklerI'm not sure currently how this is meant to be set up for CI jobs09:52
fungigboutry: it's possible the quay proxy never worked fully and the people trying to get it going lost interest in continuing to troubleshoot it, i don't recall exactly... i think it was started by contributors who were working on tripleo, if that gives any sense of the timeframe10:55
fungifrom git history, it looks like cloudnull first added it in 2019 ( https://review.opendev.org/683240 ) and then aschultz extended it in 2020/2021 ( https://review.opendev.org/744038 ) to handle a change in redirects11:00
fungiso maybe it worked initially, or maybe it's still working but tripleo jobs did something special to access it11:02
fungigboutry: looks like kolla uses it by hard-coding the port number: https://opendev.org/openstack/kolla-ansible/src/branch/master/tests/templates/globals-default.j2#L82-L8811:06
gboutrythank you very much for these precisions11:06
opendevreviewdaniel.pawlik proposed openstack/ci-log-processing master: Add url subpath for Logsender to reach OpenSearch  https://review.opendev.org/c/openstack/ci-log-processing/+/92671711:20
*** bauzas_ is now known as bauzas14:23
gboutryHow can I find how ubuntu-jammy vms are configured? I'm currently facing an issue that could be caused by CGroup configuration, and I would like to check that14:36
fungigboutry: you could adjust your job to collect the files that configure that or run some additional command to echo the information you're looking for. you can also download the images we use if you want to boot one yourself14:54
fungiwe also publish the diskimage-builder logs for them, if that helps, and you can inspect the dib elements14:55
fungigboutry: you can find samples of our current qcow2 images here: https://nb01.opendev.org/images/15:01
fungialso the dib logs are here: https://nb01.opendev.org/15:02
opendevreviewJames Page proposed openstack/project-config master: Retire the mistral charm repository  https://review.opendev.org/c/openstack/project-config/+/92679115:09
clarkbwe don't do anything special to jammy cgroups though. And yes they don't play nice with podman by default iirc15:12
clarkbyou have to tell podman to use the old cgroups system or something.15:13
clarkbbut we're about to switch the default node type to noble and I think things are better there15:13
gboutryThanks for the info15:14
gboutryI might try with noble nodeset15:14
fungiwhich has happened literally just now15:26
fungithe default nodeset change that is15:27
gboutryWell, noble solved my problems :)15:31
rosmaitai have a general question about the gerrit interface, hoping someone has a suggestion15:36
rosmaitawhen the labels with functions got migrated to submit-requirements,15:36
rosmaitathe Review-Priority field for cinder became kind of useless for us15:36
rosmaitawhat i mean is, if you look at this list of reviews,15:36
rosmaitahttps://review.opendev.org/q/project:openstack/os-brick+branch:master15:36
rosmaitathey all have a green checkmark in the "RP" column15:36
rosmaitaand setting the Review-Priority to +2, as on15:36
rosmaitahttps://review.opendev.org/c/openstack/os-brick/+/92629315:36
rosmaitadoesn't affect what shows up in the "RP" column15:36
rosmaitawhat we want is the +1 and +2 reviews to really stand out from the rest15:36
rosmaitajeremy suggested that maybe we could use hashtags15:36
rosmaitathough we would want to restrict who could assign them,15:36
rosmaitaand that might be bad if other projects use hashtags for unrestricted purposes15:36
rosmaitaanyway, i'm hoping someone might have an idea or knows of a project15:36
rosmaitathat's been able to do this successfully15:36
fungigboutry: oh! i wonder if you were rediscovering https://bugs.launchpad.net/ubuntu/+source/libpod/+bug/202439415:37
fungithat did indeed cause us issues in places too15:37
clarkbrosmaita: the issue on the change list screen is that it is showing you whether or not merge requirements have been satisfied. Since review-priority doesn't prevent merging it is satisfied by default. You can change that by requiring review-priority to be set for merging which is likely annoying for other reasons15:38
gboutryfungi: I don't think so, I think it fail at creating the root cgroup, it was failing while trying to bootstrap a single k8s15:38
clarkbre hashtags, yes I think we'd like to open those up for all projects to all registered users by default15:38
clarkbrosmaita: I think my suggestion would be to use hashtags and talk to people who abuse them. Or use a dashboard that filters by review-priority vote values15:39
gboutrykubelet exited with: failed to initialize top level QOS containers: root container [kubepods] doesn't exist # which, after research, I found happens when there's issue between cgroupv1 and v215:40
clarkbgboutry: yes, that affect jammy out of the box aiui. And it breaks podman too15:40
fungirosmaita: i think the way the release team dealt with that for their ptl-approved column was to make it a required vote, but also it really only has two possible values (0..+1)15:40
gboutrymaybe the cloud image I used locally was cgroupv2 only, hence why it did not have issues15:41
rosmaitaclarkb: is there any way to tie the Review-Priority field to the Workflow, such that if someone +W a patch, the review priority would also be satisfied?15:41
clarkbrosmaita: maybe? the language for submit requirements is fairly expressive but I'm not sure if there is a way to refer to other votes on the same change15:41
rosmaitaok, i will dig into the docs a bit15:42
gboutryfungi: I think I heard in the past there were talk about increasing the default vm size from 8gb to 16gb, is there news?15:43
opendevreviewMerged openstack/project-config master: Update update-constraints job and script  https://review.opendev.org/c/openstack/project-config/+/92640515:44
clarkbrosmaita: looking at the docs I think gerrit infers whether or not a submit-requirement is related to a label  if the label is used in hte submit requirement rule. Which would make a rule like that associate to code-review and not review-priority I think. Still may be worth trying15:45
rosmaitaclarkb: thanks15:46
fungigboutry: we have node labels for higher ram flavors in some clouds if specific jobs need those (though their availability is limited to a subset of our providers so they may have allocation problems when there are outages), but no i don't think the default is likely to change soon15:48
clarkbrosmaita: also I think if you remove the submit-requirement then that column goes away on the change list (it only shows labels required for merging)15:53
rosmaitaarrrrrgh15:55
fungibut it can still be queried15:55
clarkbessentially that functionality in gerrit is geared towards showing you what you need to do to merge15:55
clarkbit isn't there to prioritize things or sort them15:55
fungiso you can have a dashboard which lists your rp+2 changes, your rp+1 changes, etc15:55
clarkbfungi: yup thats why I suggested a dashboard15:55
fungii'll note that the lack of flexibility around this sort of thing (though it has gotten better over the years) is a big part of why the reviewday interface used to exist15:57
opendevreviewClark Boylan proposed openstack/project-config master: Use yaml.safe_load in release jobs  https://review.opendev.org/c/openstack/project-config/+/92682116:37
opendevreviewClark Boylan proposed openstack/project-config master: Bump the version of hacking to latest  https://review.opendev.org/c/openstack/project-config/+/92682517:02
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682517:17
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682517:31
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682517:40
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682518:03
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682518:19
*** bauzas_ is now known as bauzas18:30
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682518:30
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682518:42
opendevreviewClark Boylan proposed openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682518:53
opendevreviewClark Boylan proposed openstack/project-config master: Run openstack-zuul-jobs-linters on Jammy  https://review.opendev.org/c/openstack/project-config/+/92684318:53
opendevreviewMerged openstack/project-config master: Run openstack-zuul-jobs-linters on Jammy  https://review.opendev.org/c/openstack/project-config/+/92684318:55
opendevreviewClark Boylan proposed openstack/project-config master: Use yaml.safe_load in release jobs  https://review.opendev.org/c/openstack/project-config/+/92682119:24
opendevreviewClark Boylan proposed openstack/project-config master: Run ansible-lint on Ubuntu Noble  https://review.opendev.org/c/openstack/project-config/+/92684819:26
opendevreviewMerged openstack/project-config master: Omnibus fixes for running jobs on Noble with Python3.12  https://review.opendev.org/c/openstack/project-config/+/92682519:39
opendevreviewMerged openstack/project-config master: Use yaml.safe_load in release jobs  https://review.opendev.org/c/openstack/project-config/+/92682119:42
opendevreviewClark Boylan proposed openstack/project-config master: Run ansible-lint on Ubuntu Noble  https://review.opendev.org/c/openstack/project-config/+/92684819:50
opendevreviewClark Boylan proposed openstack/project-config master: Run ansible-lint on Ubuntu Noble  https://review.opendev.org/c/openstack/project-config/+/92684819:56

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!