| ianw | the only time that tag is added to the haproxy image is @ https://zuul.opendev.org/t/openstack/build/36962d5469b34d5db415aef819d9fd33/console#4/0/13/ubuntu-jammy | 00:42 |
|---|---|---|
| ianw | there is only an artifact for :latest -> https://zuul.opendev.org/t/openstack/build/36962d5469b34d5db415aef819d9fd33/artifacts | 00:43 |
| ianw | it's not retagged in the promote job -> https://zuul.opendev.org/t/openstack/build/ac9c34359ac54e60b548f5fd45f6c052/console | 00:44 |
| ianw | "skopeo copy docker://insecure-ci-registry.opendev.org:5000/quay.io/opendevorg/haproxy-statsd:36962d5469b34d5db415aef819d9fd33_latest docker://quay.io/opendevorg/haproxy-statsd:latest" must somehow also copy that tag? | 00:45 |
| ianw | skopeo copy docker://insecure-ci-registry.opendev.org:5000/quay.io/opendevorg/haproxy-statsd:36962d5469b34d5db415aef819d9fd33_latest oci:foo | 00:53 |
| ianw | and inspecting foo doesn't show that tag anywhere ... | 00:53 |
| ianw | in other words i'm at quite a loss as how it got there :/ | 00:53 |
| ianw | ok, i see it, we're still running the upload @ https://zuul.opendev.org/t/openstack/build/36962d5469b34d5db415aef819d9fd33/console | 01:08 |
| ianw | clarkb: ^ yeah, in this model, we should only run build in the gate, not upload. build will push it to the IR, and then the promote step will do the full skopeo copy | 01:13 |
| Clark[m] | Aha. So the job set is wrong. I can work on cleaning that up tomorrow. Thank you for running that down | 01:16 |
| opendevreview | Merged openstack/project-config master: Retire puppet-rally - Step 5: Remove Project https://review.opendev.org/c/openstack/project-config/+/880019 | 01:18 |
| opendevreview | Ian Wienand proposed opendev/system-config master: Don't use upload jobs for container images https://review.opendev.org/c/opendev/system-config/+/882628 | 02:00 |
| ianw | clarkb: ^ a first draft ... | 02:02 |
| *** amoralej|off is now known as amoralej | 06:27 | |
| opendevreview | Thierry Carrez proposed openstack/project-config master: Adjust ACL for ovn-bgp-agent releases https://review.opendev.org/c/openstack/project-config/+/882676 | 08:24 |
| *** diablo_rojo is now known as Guest924 | 09:16 | |
| *** amoralej is now known as amoralej|lunch | 12:13 | |
| *** dhill is now known as Guest932 | 12:21 | |
| opendevreview | Jeremy Stanley proposed openstack/project-config master: Adjust ACL for ovn-bgp-agent releases https://review.opendev.org/c/openstack/project-config/+/882676 | 12:21 |
| opendevreview | Merged openstack/project-config master: Adjust ACL for ovn-bgp-agent releases https://review.opendev.org/c/openstack/project-config/+/882676 | 12:45 |
| *** amoralej|lunch is now known as amoralej | 12:58 | |
| corvus | ianw: clarkb or should we change the upload role to skip the tags when we use the intermediate-registry method? | 13:12 |
| Clark[m] | corvus: ianw: I haven't looked too closely yet (and can't for a bit) but one upside to corvus' suggestion is we won't have to redo all the job dependency stuff. But I'm not sure what that would entail as I haven't looked at the upload job yet | 13:32 |
| corvus | i haven't looked yet either... and i'm on the fence about it: pro: it keeps the idea of being able to switch modes by flipping a variable; con: the upload job has a misleading name. | 13:40 |
| corvus | i think i'm like 70% in favor of it though. | 13:41 |
| opendevreview | Grzegorz Grasza proposed openstack/project-config master: Migrate Barbican back to Launchpad https://review.opendev.org/c/openstack/project-config/+/882716 | 13:50 |
| opendevreview | Merged openstack/project-config master: Migrate Barbican back to Launchpad https://review.opendev.org/c/openstack/project-config/+/882716 | 14:15 |
| clarkb | I failed to send out the meeting agenda yesterday. Fixing that now | 15:16 |
| fungi | thanks! | 15:17 |
| *** dmellado1 is now known as dmellado | 15:25 | |
| opendevreview | Clark Boylan proposed opendev/system-config master: Don't use upload jobs for container images https://review.opendev.org/c/opendev/system-config/+/882628 | 15:40 |
| clarkb | corvus: ianw ^ that should make ianw's approach mergable I think | 15:41 |
| clarkb | I'll look into doing the skip if intermediate registry is promote method change next | 15:41 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 15:47 |
| clarkb | corvus: ianw ^ I think that implements corvus' suggestion | 15:47 |
| opendevreview | Ching Kuo proposed opendev/system-config master: Build eavesdrop with Python 3.11 Base Images https://review.opendev.org/c/opendev/system-config/+/882564 | 15:51 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Don't use upload jobs for container images https://review.opendev.org/c/opendev/system-config/+/882628 | 15:51 |
| opendevreview | Ching Kuo proposed opendev/system-config master: Build ircbot with Python 3.11 Base Images https://review.opendev.org/c/opendev/system-config/+/881900 | 15:52 |
| clarkb | If we go with 882628 we can use the grafyaml move as a test case | 15:55 |
| clarkb | and not rebuild everything else | 15:56 |
| clarkb | er sorry I meant 882724 | 15:56 |
| clarkb | gmann: sorry about the need for rebases... Once I get through the quay.io migration I'll try to get your changes sorted out | 15:56 |
| clarkb | bah I'm having a morning. That was for genekuo not gmann (sorry gmann) | 15:57 |
| *** amoralej is now known as amoralej|off | 15:59 | |
| gmann | clarkb: ok, no worry :) | 16:02 |
| genekuo | no problem, I can wait for the quay.io changes to go through | 16:09 |
| clarkb | genekuo: thank you for the patience and the help! | 16:10 |
| opendevreview | Clark Boylan proposed opendev/system-config master: Don't use upload jobs for container images https://review.opendev.org/c/opendev/system-config/+/882628 | 16:11 |
| corvus | clarkb: am i sleep deprived or is the logic backwards on 882724? | 16:14 |
| corvus | clarkb: that looks like "when method=intermediate-registry then include_tasks: push.yaml" | 16:15 |
| clarkb | oh yup thats backwards | 16:15 |
| corvus | okay whew, so probably just needs a not in front of the whole condition | 16:16 |
| clarkb | when: promote_container_image_method is not defined or promote_container_image_method != 'intermediate-registry' <- that should do it? | 16:18 |
| clarkb | if we are not defined then push. If we are defined and value is tag then push. If we are defined and value is intermediate-registry don't push | 16:18 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 16:19 |
| corvus | looks like we are treating the default as tag... and actually when: promote_container_image_method|default('tag') != 'intermediate-registry' would be consistent with the use in the promote role | 16:20 |
| corvus | maybe even | 16:21 |
| corvus | when: promote_container_image_method|default('tag') == 'tag' | 16:21 |
| clarkb | ++ I'll update | 16:21 |
| corvus | clarkb: oh i think we also need to move or copy that role var documentation.. it's currently only in the promote role | 16:22 |
| corvus | so we should either copy it to the upload role or put in in the common text | 16:22 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 16:35 |
| clarkb | corvus: ^ like that maybe? | 16:35 |
| corvus | clarkb: ++ good call on the promote switch too | 16:44 |
| clarkb | " Don't compare to literal True/False." | 17:01 |
| clarkb | I feel like this is an instance where it is more readable... | 17:01 |
| opendevreview | Jeremy Stanley proposed opendev/zone-opendev.org master: Dummy mailman hostname to house the list of lists https://review.opendev.org/c/opendev/zone-opendev.org/+/867981 | 17:01 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 17:02 |
| clarkb | corvus: ^ that should make ansible-lint happier | 17:02 |
| clarkb | yup zuul is happy with that change now | 17:23 |
| fungi | new held lists99 node is 23.253.160.97 and in its initial state the postorius and hyperkitty interfaces are correctly filtering lists by domain, the only problem is that hyperkitty says "lists.opendev.org" in the top-left corner of all pages even for lists.zuul-ci.org archives (hyperlinks also still go to the right places) | 18:44 |
| clarkb | fungi: I seem to recall that is a singular value it pulls out in the template | 18:50 |
| clarkb | but you can double check that | 18:50 |
| fungi | yeah, i'm back to poking around at the django admin interface, we seem to still be creating an initial "lists.opendec.org" django site even though i thought https://review.opendev.org/867986 would make it mailman.opendev.org so clearly there's more to it than just being the first site we create from the mailman_sites list in hostvars | 18:59 |
| fungi | i guess it's the SERVE_FROM_DOMAIN at https://opendev.org/opendev/system-config/src/branch/master/playbooks/roles/mailman3/templates/docker-compose.yaml.j2#L43 | 19:03 |
| fungi | aha, there's a default domain called "example.com" and then it gets renamed here to whatever's in that value: https://opendev.org/opendev/system-config/src/branch/master/docker/mailman/web/docker-entrypoint.sh#L139-L146 | 19:06 |
| clarkb | ya and it is a single value... so we may just have to accept this for now | 19:07 |
| fungi | though the other magic with that is happening in the settings here: https://opendev.org/opendev/system-config/src/branch/master/playbooks/roles/mailman3/files/web-settings.py#L227-L230 | 19:10 |
| opendevreview | Jeremy Stanley proposed openstack/project-config master: Add a convenient normalize-gerrit-acl tox testenv https://review.opendev.org/c/openstack/project-config/+/882750 | 19:36 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Separate mailman default hostname https://review.opendev.org/c/opendev/system-config/+/867986 | 19:48 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Use magic domain guessing in Mailman 3 https://review.opendev.org/c/opendev/system-config/+/867987 | 19:48 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: DNM force mm3 failure to hold the node https://review.opendev.org/c/opendev/system-config/+/855292 | 19:48 |
| opendevreview | Jeremy Stanley proposed opendev/system-config master: Upgrade to latest Mailman 3 releases https://review.opendev.org/c/opendev/system-config/+/869210 | 19:48 |
| fungi | okay, that stack is updated for the SERVE_FROM_DOMAIN adjustment so i can see where things wind up, new autohold set too | 19:51 |
| clarkb | I'm going to pop out for a long lunch/exercise break. I'll be back to hopefully push quay things along a bit further | 19:51 |
| ianw | i'm uneasy on https://review.opendev.org/c/zuul/zuul-jobs/+/882724. i've made a few comments and added a link to other docs i think probably want updating if that's the path we go on. i've always had in my head it's "check - build, gate - build, promote - copy from IR" in this model | 20:30 |
| ianw | ... and that's how i wrote some bits of the docs i contributed too. so i'm having a hard time shifting that view :) but i'm not -1 if others disagree | 20:31 |
| clarkb | ianw: did the updates I made to your system-config change to make it pass look good to you? In particular i moved everything back into a common vars anchor because the promote job needs to know about the image details too | 22:00 |
| clarkb | corvus: I'm happy to update the other repos that I did the upload job in to only use build. I think it mostly comes down to your opinion on the best thing for zuul | 22:00 |
| clarkb | I also understand ianw's argument that this is far more explicit and probably easier to understand. The reason I got it wrong was the conversion path made it confusing but if starting from scratch this probably does make more sense to a new implementation | 22:01 |
| ianw | clarkb: yeah, i did think the promote job just promoted the artifact in the job it was pointed at | 22:06 |
| ianw | i must have read it wrong | 22:06 |
| ianw | oh i see now; https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/promote-container-image/tasks/promote-from-intermediate-registry.yaml#L18 | 22:08 |
| clarkb | ianw: it may also create the repo in quay if necessary. And i think the push also needs the info as well | 22:08 |
| ianw | sorry yeah, that was me misreading that | 22:09 |
| ianw | 882628 lgtm. but i don't want to be the initiator of a bunch of work to reorg the jobs for purely theoretical reasons, as we found in the matrix channel it's unlikely anyone's even read the updated docs :) | 22:11 |
| fungi | 173.231.255.71 is the new-new-new-new-... held lists99 | 22:18 |
| ianw | Re: Fw: latest: Re: final: Fw: latestv2 | 22:23 |
| clarkb | aroo? | 22:34 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 22:41 |
| clarkb | ianw: ^ I tried to address your comments in that patchset | 22:41 |
| corvus | clarkb: ianw i think ianw is right that if we go with my approach, we should probably rework the documentation to change the recommendation (so it says always run "upload" in "gate"). | 22:45 |
| clarkb | ok that last patchset said it was optional. But I can update it to be the only suggestions | 22:46 |
| corvus | but i also don't feel very strongly about this, and if we prefer to change out the jobs, that's fine. i think i'm getting cognitive dissonance though because we explicitly decided *not* to have two different promote roles/jobs for the different tasks, but we are considering having two different "gate" jobs | 22:47 |
| corvus | i don't have any objection to ianw's change. i only thought i was making a suggestion that maybe was aligned with our more recent decisions | 22:48 |
| clarkb | fwiw I think zuul also did build + upload with promote method intermediate-registry so ya I think at the point of adding these jobs to repos we ended up with what was familiar possibly entirely unintentionally (it was in my case) | 22:50 |
| clarkb | I agree that build + build was the previous intention, but maybe our unintentional mistake is an indicator the consistent thing is better for users? | 22:51 |
| corvus | yeah i think you put your finger on it | 22:51 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 22:52 |
| clarkb | corvus: ^ that update says only build + upload | 22:52 |
| clarkb | so if we go that route it is consistent with a single suggestion | 22:52 |
| clarkb | it is late enough in my day that it may be worth sleeping on it too and making a decision tomorrow | 22:53 |
| corvus | wfm | 22:54 |
| ianw | i'm happy to go with the -upload role, especially with the updates to 882724 that lgtm, modulo some small comments i just left | 23:15 |
| opendevreview | Clark Boylan proposed zuul/zuul-jobs master: Gracefully handle use of intermediate registry in container upload role https://review.opendev.org/c/zuul/zuul-jobs/+/882724 | 23:18 |
| clarkb | ianw: thank you for the reviews | 23:18 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!