| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Add a zuul-ensure-passwords role https://review.opendev.org/717880 | 00:06 |
|---|---|---|
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Add a zuul-ensure-passwords role https://review.opendev.org/717880 | 00:10 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add bionic-plain non-voting testing https://review.opendev.org/717655 | 00:14 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: test jobs: fixup autogeneration header https://review.opendev.org/717656 | 00:14 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 00:14 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Add zuul-registry deployment https://review.opendev.org/710650 | 00:14 |
| *** rlandy|afk is now known as rlandy | 00:22 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add bionic-plain non-voting testing https://review.opendev.org/717655 | 00:25 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: test jobs: fixup autogeneration header https://review.opendev.org/717656 | 00:25 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 00:25 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 00:27 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 00:27 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 00:28 |
| tristanC | corvus: it seems like we are still missing https://review.opendev.org/708860 to get the zuul-operator image actually published | 00:29 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Add bionic-plain non-voting testing https://review.opendev.org/717655 | 00:35 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: test jobs: fixup autogeneration header https://review.opendev.org/717656 | 00:35 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 00:35 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 00:37 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 00:38 |
| openstackgerrit | Clark Boylan proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 00:39 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Add bionic-plain non-voting testing https://review.opendev.org/717655 | 00:52 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: test jobs: fixup autogeneration header https://review.opendev.org/717656 | 00:52 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 00:52 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 00:52 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 00:52 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 00:52 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 01:08 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 01:08 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 01:08 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 01:08 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 01:15 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 01:15 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 01:15 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 01:15 |
| *** swest has quit IRC | 01:28 | |
| *** zxiiro has quit IRC | 01:39 | |
| *** swest has joined #zuul | 01:43 | |
| *** cdearborn has quit IRC | 01:43 | |
| *** Goneri has quit IRC | 02:00 | |
| *** ysandeep|away is now known as ysandeep|rover | 02:02 | |
| *** igordc has joined #zuul | 02:33 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 02:59 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 02:59 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 03:00 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 03:00 |
| *** ysandeep|rover is now known as ysandeep|afk | 03:13 | |
| *** igordc has quit IRC | 03:16 | |
| *** bhavikdbavishi has joined #zuul | 03:20 | |
| *** bhavikdbavishi1 has joined #zuul | 03:22 | |
| *** bhavikdbavishi has quit IRC | 03:24 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:24 | |
| *** bhavikdbavishi has quit IRC | 03:36 | |
| *** rlandy has quit IRC | 03:37 | |
| *** bhavikdbavishi has joined #zuul | 03:38 | |
| *** bhavikdbavishi has quit IRC | 03:45 | |
| *** bhavikdbavishi has joined #zuul | 04:08 | |
| *** ysandeep|afk is now known as ysandeep|rover | 04:18 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 04:21 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 04:21 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 04:21 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 04:22 |
| *** evrardjp has quit IRC | 04:36 | |
| *** evrardjp has joined #zuul | 04:37 | |
| openstackgerrit | Andreas Jaeger proposed zuul/zuul-jobs master: Update registry test to use ensure-podman and ensure-docker https://review.opendev.org/716752 | 04:53 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 04:53 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: fetch-zuul-cloner: prefer venv to virtualenv https://review.opendev.org/717882 | 04:53 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 04:53 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 04:53 |
| openstackgerrit | Merged zuul/zuul-jobs master: Update registry test to use ensure-podman and ensure-docker https://review.opendev.org/716752 | 05:14 |
| *** saneax has joined #zuul | 05:20 | |
| *** y2kenny has quit IRC | 05:32 | |
| *** dpawlik has joined #zuul | 06:30 | |
| *** avass has quit IRC | 06:53 | |
| *** avass has joined #zuul | 06:53 | |
| *** gtema has joined #zuul | 06:55 | |
| *** rpittau|afk is now known as rpittau | 06:56 | |
| *** pleia2 has quit IRC | 07:02 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Retry nodeenv creation https://review.opendev.org/717820 | 07:04 |
| *** pleia2 has joined #zuul | 07:04 | |
| *** bhavikdbavishi has quit IRC | 07:11 | |
| *** sgw has quit IRC | 07:11 | |
| *** jcapitao has joined #zuul | 07:18 | |
| *** tosky has joined #zuul | 07:24 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Don't try to stream from winrm connections https://review.opendev.org/717973 | 07:38 |
| *** ysandeep|rover is now known as ysandeep|lunch | 07:44 | |
| tobiash | tristanC: I've added a question on https://review.opendev.org/708860 | 08:08 |
| openstackgerrit | Merged zuul/zuul-jobs master: docs: fix a typo in `run-test-command` https://review.opendev.org/717713 | 08:15 |
| *** bhavikdbavishi has joined #zuul | 08:33 | |
| *** ysandeep|lunch is now known as ysandeep|rover | 08:44 | |
| *** sshnaidm|afk is now known as sshnaidm | 08:57 | |
| openstackgerrit | Merged zuul/zuul master: Render buildset progress bar correctly https://review.opendev.org/716878 | 09:08 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Support ssh-enabled windows hosts in add-build-sshkey https://review.opendev.org/653712 | 09:12 |
| zbr | tristanC: tobiash: can you please recheck https://review.opendev.org/#/c/690057/ -- i really need to be able to upgrade tox | 09:27 |
| zbr | on ansible zuul instance there is an outdated version pre-installed, and this would enable me the bump it from consuming jobs. | 09:28 |
| * tobias-urdin just passed 10k zuul builds since moving away from jenkins | 10:01 | |
| *** rpittau is now known as rpittau|bbl | 10:25 | |
| tobiash | tobias-urdin: congratulations :) | 10:59 |
| *** bhavikdbavishi has quit IRC | 11:06 | |
| *** bhavikdbavishi has joined #zuul | 11:11 | |
| tobiash | zbr: you already got a comment | 11:13 |
| tobiash | however I don't understand it | 11:14 |
| zbr | yep, addressing it. minor stuff | 11:14 |
| tobiash | k | 11:14 |
| zbr | maybe someone knows how to avoid repeating the same uninstall task 3 times in the file | 11:14 |
| *** jcapitao is now known as jcapitao_lunch | 11:19 | |
| *** gtema has quit IRC | 11:22 | |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: pagure: Improve CI status flag handling https://review.opendev.org/718063 | 11:22 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 11:24 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Add zuul-operator-upload-image job https://review.opendev.org/708860 | 11:27 |
| *** ysandeep|rover is now known as ysandeep|afk | 11:29 | |
| *** hashar has joined #zuul | 11:31 | |
| *** hashar has quit IRC | 11:42 | |
| tristanC | zbr: why not using include_tasks? | 11:48 |
| zbr | tristanC: not a big fun of them, especially for a single task. | 11:48 |
| zbr | but i may have to, i am still unsure if looping with become works or not | 11:48 |
| tristanC | zbr: the `Verify tox is not installed` task is also repeated | 11:49 |
| zbr | not fully, the last verification is quite different | 11:50 |
| tristanC | zbr: i find the yaml anchor and merge operator rather confusing | 11:50 |
| zbr | tristanC: nopb, i can avoid it. | 11:50 |
| zbr | there is also another way of doing it, adding remove support to the role itself. | 11:51 |
| zbr | i could use tasks_from: remove.yaml and keep the removal code inside the role itself, instead of having it part of the test playbook. | 11:52 |
| zbr | some could find it useful | 11:52 |
| tristanC | zbr: that sounds like a bad idea. and how would you test that without introducing more repetition? | 11:52 |
| zbr | i would just run removal, and assume it worked. | 11:53 |
| zbr | tristanC: if I add a removal tasks file, only for the test playbook, where should I put it? how to name it? | 11:55 |
| zbr | like test-playbooks//tasks/tox-remove.yaml ? | 11:55 |
| tristanC | zbr: i meant, adding a remove task to an install role just to avoid the repetition of a task in the zuul-test playbook sounds odd | 11:56 |
| *** ysandeep|afk is now known as ysandeep|rover | 11:56 | |
| zbr | i do have an use-case: if you want to enforce removal of tox somewhere. but I will do the local implementation for the moment and not touch the role | 11:57 |
| *** sshnaidm is now known as sshnaidm|afk | 11:58 | |
| tristanC | zbr: if we need install-* role to be able to uninstall, other than for install-* testing purpose, then that's another story. | 11:59 |
| zbr | install is no more, now we have ensure | 11:59 |
| tristanC | s/install-/ensure-/ | 12:00 |
| zbr | and i can see a (low) use case for "ensure foo is missing" ;) | 12:00 |
| tristanC | zbr: the semantic looks odd, why would you use an ensure-foo role to remove foo? | 12:02 |
| zbr | tristanC: it is a common pattern on configuration management, that is why ansible does have only one package module and it does not have a "install-package" and "remove-package" module. | 12:03 |
| zbr | state determines the action, defaults to present. | 12:03 |
| zbr | anyway, i am not going to propose that, at least not now. | 12:03 |
| *** gtema has joined #zuul | 12:04 | |
| tristanC | zbr: in ansible module that make sense as the semantic is `package.state = <absent|present>` . Using `package` doesn't imply the state to be present. In zuul-jobs we have `ensure-foo` which already imply that foo state will be present. | 12:05 |
| *** armstrongs has joined #zuul | 12:06 | |
| *** rpittau|bbl is now known as rpittau | 12:06 | |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: pagure: Improve CI status flag handling https://review.opendev.org/718063 | 12:08 |
| *** armstrongs26 has joined #zuul | 12:11 | |
| zbr | tristanC: https://opendev.org/zuul/zuul-jobs/src/branch/master/test-playbooks/ensure-tox.yaml#L34-L57 -- this seems like a bug, running removing twice. | 12:13 |
| zbr | i see no reason to run twice | 12:13 |
| armstrongs26 | hey have a question. We have a common framework untrusted repo that is defined in around 16 tenants. It is used a dependency in all these tenants. We noticed when we commit to master to change this dependant project it causes a queue on the zuul scheduler. Is there a way of setting this up to avoid the queue delay? | 12:14 |
| *** armstrongs has quit IRC | 12:14 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 12:18 |
| *** armstrongs has joined #zuul | 12:19 | |
| *** rfolco has joined #zuul | 12:20 | |
| *** rlandy has joined #zuul | 12:24 | |
| *** jcapitao_lunch is now known as jcapitao | 12:31 | |
| openstackgerrit | Albin Vass proposed zuul/zuul master: WIP: Enables whitelisting and configuring callbacks https://review.opendev.org/717260 | 12:31 |
| *** hashar has joined #zuul | 12:31 | |
| *** armstrongs has quit IRC | 12:32 | |
| *** igordc has joined #zuul | 12:34 | |
| mnaser | armstrongs26: are you saying that it gets queued in all of the tenants? | 12:52 |
| openstackgerrit | Graham Hayes proposed zuul/nodepool master: Implement an Azure driver https://review.opendev.org/554432 | 13:05 |
| *** bhavikdbavishi has quit IRC | 13:07 | |
| armstrongs26 | mnaser: we see the scheduler pause scheduling jobs for a couple of minutes and the build queue has a big number of management events. When it goes back down jobs start queuing again | 13:07 |
| armstrongs26 | when i say pause i mean new requests queue and wait and don't get scheduled by nodepool | 13:08 |
| *** sgw has joined #zuul | 13:12 | |
| openstackgerrit | Albin Vass proposed zuul/zuul master: Rename getAnsiblePluginDir to getZuulAnsiblePluginDir https://review.opendev.org/718102 | 13:18 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: Rename getAnsiblePluginDir to getZuulAnsiblePluginDir https://review.opendev.org/718102 | 13:18 |
| *** Goneri has joined #zuul | 13:23 | |
| *** sshnaidm|afk is now known as sshnaidm | 13:27 | |
| openstackgerrit | Albin Vass proposed zuul/zuul master: WIP: Enables whitelisting and configuring callbacks https://review.opendev.org/717260 | 13:28 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: WIP: Enables whitelisting and configuring callbacks https://review.opendev.org/717260 | 13:29 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: Rename getAnsiblePluginDir to getZuulAnsiblePluginDir https://review.opendev.org/718102 | 13:33 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: Rename getAnsibleDir to getZuulAnsibleDir https://review.opendev.org/718102 | 13:35 |
| mnaser | in cool news, cheroot and cherrypy is getting testing via zuul, though no gating yet | 13:36 |
| *** gtema has quit IRC | 13:45 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Log missing required status checks https://review.opendev.org/718114 | 13:47 |
| mordred | mnaser: \o/ | 13:48 |
| *** harrymichal has joined #zuul | 13:50 | |
| tobiash | zbr: I wonder if installing tox into a venv and linking it into PATH would be easier than this uninstall dance | 13:51 |
| tobiash | armstrongs26: a merge to master requires a tenant reconfiguration in all tenants that use this repo | 13:54 |
| tobiash | armstrongs26: currently there is no way around this but with the ha scheduler work we'll be switching from a single event queue to a distributed multi-worker event queue | 13:55 |
| tobiash | armstrongs26: this won't then block zuul globally in this case | 13:55 |
| armstrongs26 | was just checking to see if there is any better way for configuring common projects. Thanks for the info. Looking forward to the distributed scheduler :) | 13:57 |
| *** avass has quit IRC | 14:08 | |
| *** avass has joined #zuul | 14:10 | |
| *** gtema has joined #zuul | 14:12 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Add bionic-plain non-voting testing https://review.opendev.org/717655 | 14:14 |
| *** gtema has quit IRC | 14:25 | |
| openstackgerrit | Fabien Boucher proposed zuul/zuul master: pagure: Make use of the new project webhook/token endpoint https://review.opendev.org/717732 | 14:27 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: Use ensure-* roles https://review.opendev.org/718124 | 14:27 |
| zbr | mordred: clarkb: can we merge https://review.opendev.org/#/c/716578/ ? (ensure-podman) | 14:32 |
| *** ysandeep|rover is now known as ysandeep|away | 14:33 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 14:34 |
| *** harrymichal has quit IRC | 14:35 | |
| *** bhavikdbavishi has joined #zuul | 14:44 | |
| *** bhavikdbavishi1 has joined #zuul | 14:46 | |
| *** bhavikdbavishi has quit IRC | 14:48 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 14:48 | |
| *** bhavikdbavishi has quit IRC | 15:03 | |
| *** sshnaidm is now known as sshnaidm|afk | 15:07 | |
| mnaser | mordred, corvus: so fetch-javascript-output posts an artifact called "Site preview" -- but the zuul-preview code seems to (probably) need an artifact name without a space? | 15:09 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 15:11 |
| *** bhavikdbavishi has joined #zuul | 15:11 | |
| corvus | mnaser: yeah, it's got to be something valid for dns, so only letters, numbers, '-' in the name | 15:14 |
| corvus | mnaser: maybe we should change zuul-preview to use the 'type' metadata entry instead of the name | 15:15 |
| corvus | mnaser: that's what we did with the artifact fetching roles | 15:15 |
| *** sshnaidm|afk is now known as sshnaid | 15:27 | |
| *** sshnaid is now known as sshnaidm | 15:27 | |
| mnaser | corvus: that seems reasonable, and as much as i dont want to complicate things, we need to figure out if there is a possibility of multiple previews i guess | 15:29 |
| openstackgerrit | Graham Hayes proposed zuul/nodepool master: Implement an Azure driver https://review.opendev.org/554432 | 15:30 |
| corvus | mnaser: here's the previous use of this: http://zuul.opendev.org/t/openstack/build/167715b656ee4504baa940c5bd9f3821 | 15:30 |
| corvus | mnaser: we just named the artifact "site" | 15:30 |
| mnaser | corvus: yeah, i guess i'll have to template it out inside here -- https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/fetch-javascript-output/tasks/main.yaml | 15:31 |
| mnaser | as i'm using the zuul built-in build-javascript-content job | 15:33 |
| openstackgerrit | James E. Blair proposed zuul/zuul-preview master: Look up artifacts by metadata.type rather than name https://review.opendev.org/718151 | 15:35 |
| corvus | mnaser: ^ if we want to go that way | 15:35 |
| mnaser | corvus: that looks good to me, to get a head start, ill push a change to zuul/zuul-jobs :) | 15:36 |
| corvus | mnaser: yes, we'll need to update https://opendev.org/zuul/zuul-jobs/src/branch/master/roles/fetch-javascript-output/tasks/main.yaml to either change the name or add the metadata.type field; then we'll want to return a second artifact on the build job. that shouldn't be in zuul-jobs, since it's going to have opendev site-specific information. | 15:36 |
| tobiash | corvus: we noticed that our executors get a significant performance benefit when we do a regular (e.g. weekly cache wipe). What do you think about a builtin mechanism to facilitate this in an easy way? | 15:38 |
| tobiash | our current approach is pause a bunch of executors, wait intil there is no starting build, wipe the cache and unpause again | 15:38 |
| corvus | mnaser: (we need one artifact that is the url on the log server, and we'll want a second artifact that's the link to the preview server. we'll have to figure out the best way to name those, since they are both preview links) | 15:38 |
| corvus | tobiash: we see the same improvement whenever we clean caches; it's all the throwaway refs. i think we should automate it; we probably don't have to delete the whole repo, we can probably figure out what refs to delete and then run a gc | 15:39 |
| tobiash | do they accumulate? | 15:40 |
| corvus | tobiash: if we do it that way, it can probably be very quick and maybe very frequent (maybe run a process every few hours that deletes refs older than 24 hours) | 15:40 |
| corvus | tobiash: i think so, but i haven't looked lately | 15:41 |
| corvus | tobiash: hrm, maybe we don't store refs anymore | 15:42 |
| corvus | tobiash: if that's the case, i wonder what it is... | 15:42 |
| corvus | tobiash: because git should automatically run gc | 15:42 |
| tobiash | corvus: at least for github we seem to accumulate local branches that don't have matching remote refs anymore | 15:43 |
| corvus | tobiash: ah, good point. we don't see that, but that should be easy to automate too | 15:44 |
| tobiash | corvus: gc only prunes objects older than two weeks by default so on a busy system that might still leave a lot of unneeded objects around | 15:44 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 15:45 |
| tobiash | corvus: so I guess we need to manually run git gc with a custom prune period | 15:47 |
| tobiash | and solve the branch leak for github | 15:47 |
| clarkb | zbr: looking now | 15:47 |
| zbr | thanks. i am still working on ensure-tox one, seems to be quite hard to eradicate tox from ubuntu hosts :D | 15:48 |
| clarkb | zbr: I've approved the change but elft a note on it if you want to address that in a followup | 15:49 |
| clarkb | (its a minor thing) | 15:49 |
| tobiash | corvus, tristanC: hrm, we just have been hit by NoneType has no loading_errors (https://review.opendev.org/676947) as well | 15:53 |
| openstackgerrit | Merged zuul/zuul master: Add yappi and objgraph to container image https://review.opendev.org/717602 | 15:56 |
| tristanC | corvus: tobiash: this also seems relevant: https://review.opendev.org/#/c/712544/1 , not sure what to do about it though | 15:56 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: WIP: Implement graceful termination for the executor https://review.opendev.org/718157 | 15:57 |
| mnaser | corvus: actually, maybe a good time to ask, i was a bit confused as to what the point of uploading to log server vs uploading to "preview" server? is this similar to a upload/promote type of behaviour? | 15:57 |
| tristanC | having Optional[] type annotation would help understand and enforce how missing layout are handled | 15:57 |
| mordred | mnaser: the preview server is just a proxy service that allows serving a log server location from a root url - so for things like the gatsby build that just don't work if you try to suburl them with an arbitrary suburls | 15:59 |
| mnaser | OH, okay, i get it | 15:59 |
| mnaser | so one artifact is going to be "contents" and the other is "view website" | 16:00 |
| mnaser | i was thinking of using success-url and not using an artifact | 16:00 |
| tobiash | tristanC: that looks like the same bugfix :) | 16:00 |
| *** dpawlik has quit IRC | 16:00 | |
| tobiash | we had 8 hits of that exception (and reports) in the last 7 days, so seems to be a very rare corner case or some sort of race | 16:00 |
| corvus | tobiash, tristanC: yeah, there is a bug somewhere, but i don't think anyone has a theory for what it is; it needs more investigation/debugging. the bug is not fixed by that patch, that patch would only mask it. | 16:01 |
| *** bhavikdbavishi has quit IRC | 16:01 | |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Add zuul-registry deployment https://review.opendev.org/710650 | 16:01 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: test: refactor run tasks to include file https://review.opendev.org/718158 | 16:01 |
| tobiash | I'm aware of that, digging in the logs now | 16:01 |
| corvus | i guess that was mostly for tristanC then :) | 16:02 |
| openstackgerrit | Merged zuul/zuul-jobs master: Add support for RedHat platforms on ensure-podman https://review.opendev.org/716578 | 16:03 |
| corvus | mnaser: i think we're trying to stop using success-urls, since if we use them, there's no way to actually get to the logs. so i think it should be a second artifact | 16:03 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Annotate dynamic layout creation https://review.opendev.org/718160 | 16:03 |
| *** sshnaidm is now known as sshnaidm|afk | 16:03 | |
| mnaser | corvus: fair enough, i think an artifact of type zuul-website is probably what we'd use, but yeah, there's some trickery around finding the right word | 16:04 |
| mnaser | but that's an opendev-ism, not a zuul-ism, we can probably stick with 'site' as type | 16:04 |
| tobiash | corvus: I just got the info from the user that depends-on and a yaml syntax error are involved | 16:05 |
| corvus | tobiash: i'm not having any luck making opendev repos any smaller with git gc; so maybe the main outstanding issue is the extra branches you see due to the github workflow? | 16:07 |
| corvus | maybe everything else is automatic now | 16:08 |
| *** rpittau is now known as rpittau|afk | 16:08 | |
| tobiash | corvus: probably, but maybe in gerrit case the size is not the main factor but the number of objects not in a pack | 16:08 |
| corvus | tobiash: yeah, i haven't tried more extensive repacks | 16:09 |
| tobiash | side note, don't try --aggressive ;) | 16:09 |
| tobiash | it easily ooms | 16:09 |
| openstackgerrit | Tristan Cacqueray proposed zuul/zuul-operator master: Increase scheduler wait to timeout https://review.opendev.org/718162 | 16:10 |
| tristanC | Here is a simple review to improve zuul-operator test reliability ^ | 16:11 |
| openstackgerrit | Mohammed Naser proposed zuul/zuul-jobs master: fetch-javascript-output: add metadata.type https://review.opendev.org/718163 | 16:11 |
| mnaser | ^ so that should get us the type | 16:12 |
| mnaser | so once https://review.opendev.org/#/c/718151/ lands, we'll be able to see the preview of the zuul gatsby-ified website | 16:13 |
| openstackgerrit | Merged zuul/zuul master: tox: rename pep8 to linters https://review.opendev.org/703635 | 16:15 |
| clarkb | mnaser: mordred: that is a backward incompatbile change right? any concern with that? | 16:20 |
| clarkb | we could continue to fallback to name and be backward compatbile | 16:20 |
| mordred | clarkb: I mean - proxy isn't in use because of the security issue that was just fixed | 16:22 |
| mordred | so - I think it's fairly safe to just do it | 16:22 |
| mordred | I mean - I _suppose_ someone other than us could be using it who cares less about open proxies | 16:23 |
| openstackgerrit | Merged zuul/zuul master: tox: do not install bindep for linters https://review.opendev.org/703636 | 16:24 |
| tobiash | corvus: also I found out that the repo of the parent item has a .zuul.yaml as well as a zuul.d | 16:26 |
| tobiash | but that might be unrelated | 16:29 |
| fungi | tobiash: according to https://github.com/ether/etherpad-lite/issues/3781 the zuul.d would be used and .zuul.yaml ignored | 16:32 |
| tobiash | fungi: that is correct | 16:32 |
| fungi | unless you actually meant .zuul.yaml and .zuul.d in which case .zuul.yaml is read and .zuul.d ignored | 16:32 |
| tobiash | just found that warning in the context of the issue | 16:32 |
| fungi | er, wrong link from my paste buffer, sorry | 16:33 |
| fungi | https://zuul-ci.org/docs/zuul/reference/config.html#configuration-loading | 16:33 |
| * fungi is doing too many things at once | 16:33 | |
| tobiash | fungi: code and docs are in sync for that :) | 16:34 |
| openstackgerrit | Merged zuul/zuul-jobs master: fetch-javascript-output: add metadata.type https://review.opendev.org/718163 | 16:36 |
| *** evrardjp has quit IRC | 16:36 | |
| *** evrardjp has joined #zuul | 16:37 | |
| *** harrymichal has joined #zuul | 16:37 | |
| *** zxiiro has joined #zuul | 16:39 | |
| corvus | clarkb: i think i'll approve 151 under the assumption it has no users | 16:42 |
| corvus | clarkb: if someone yells, it's certainly possible to support both, but i think it's desirable to be clear and only support one | 16:42 |
| clarkb | ok | 16:42 |
| openstackgerrit | Merged zuul/nodepool master: Use ensure-* roles https://review.opendev.org/717822 | 16:52 |
| openstackgerrit | Merged zuul/zuul-preview master: Look up artifacts by metadata.type rather than name https://review.opendev.org/718151 | 16:59 |
| openstackgerrit | Merged zuul/zuul master: Use ensure-* roles https://review.opendev.org/718124 | 17:10 |
| *** harrymichal has quit IRC | 17:11 | |
| *** harrymichal has joined #zuul | 17:11 | |
| *** harrymichal has quit IRC | 17:40 | |
| *** armstrongs26 has quit IRC | 17:45 | |
| clarkb | fungi: provides and requires are oriented around artifacts, they don't really care about projects or branches as far as I know | 17:50 |
| clarkb | fungi: a job will say I provide this artifact foo, then if you require that artifact foo any jobs providing it will run before you run | 17:51 |
| fungi | clarkb: yeah, so how would you go about ensuring that the artifact for the build of that job ahead of the current change completed? | 17:54 |
| clarkb | fungi: zuul does that for you | 17:55 |
| fungi | assuming they're both for the same project | 17:55 |
| fungi | so your current buildset has a job which provides that, but by virtue of being for the same project+branch the change ahead of that one in the queue also provides that from the build of the same job in its buildset | 17:55 |
| fungi | and it sounded like what he was looking for was a way to be sure that the artifact provided by the build in the change ahead of the current one in the queue had completed | 17:56 |
| clarkb | yup | 17:56 |
| fungi | so... how does zuul differentiate then? | 17:56 |
| clarkb | thats how the docker image jobs are intended to work aiui and they rely on provides, requires to do that | 17:56 |
| clarkb | fungi: because provides and requires are between buildsets and dependencies are within a buildset | 17:57 |
| clarkb | so you've got two constructs and you have to configure both | 17:57 |
| clarkb | in the docker image case you both provide and requires the same artifact | 17:57 |
| clarkb | that way you can wait for parent change to complete that build then update the image the parent provided | 17:58 |
| clarkb | then in the buildset you depend on your job to build that image in order to consume it for testing | 17:58 |
| fungi | so change 1 is approved and enqueued for project foo and runs a job which declares that it provides artifact bar. change 2 for foo is approved and enqueued into the same dependent pipeline and also runs the same job which claims to provide artifact bar. a second job for change 2 requires artifact bar from change 1's buildset... how? | 17:58 |
| *** jcapitao has quit IRC | 17:59 | |
| clarkb | the job providing artifact bar in change 2 doesn't start until the one in change 1 completes. This is configured via the "requires" directive. This allows change 2's artifact production to incorporate what change 1 has done | 18:00 |
| clarkb | the second job in change 2 does not directly depend on the artifact in change 1, instead it depends on the artifact produced by change 2, but if that artifact production requires the artifact from change 1 you get the ordering right | 18:01 |
| fungi | i see, so you can chain then by saying don't start the job which produces artifact bar in this change until all the buildsets enqueued ahead have also already produced artifact bar? | 18:01 |
| clarkb | yes | 18:01 |
| fungi | and you express that by saying this job provides bar and requires bar (on the same job)? or is it implicit that zuul won't run two builds in different buildsets concurrently if they both claim to provide the same artifact? | 18:03 |
| clarkb | you have to provides and requires on the same job pretty sure this is what the docker image builds do | 18:03 |
| fungi | okay | 18:03 |
| *** harrymichal has joined #zuul | 18:06 | |
| *** bhavikdbavishi has joined #zuul | 18:07 | |
| mordred | clarkb, fungi : almost but not quite ... you would not have a single job providing and requiring the same artifact | 18:09 |
| *** hashar has quit IRC | 18:10 | |
| mordred | fungi: in your example above, the second job in change 2 that requires artifact bar isn't going to be looking for bar from change 1's buildset, it's going to be looking for it from change 2's buildset, since change 2 is building it and providing it | 18:11 |
| mordred | however, if you uploaded change 2 and it _didn't_ trigger artifact bar to be built, then the second job that requires artifact bar would get the artifact from change 1's build rather than change 2's | 18:12 |
| fungi | i'll admit i was reading between the lines for the question in the ml, but it sounded like the desire was to be able to compare the artifact built in one change against the same artifact enqueued ahead of it, to see how they differ | 18:12 |
| *** bhavikdbavishi1 has joined #zuul | 18:12 | |
| *** bhavikdbavishi has quit IRC | 18:12 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 18:12 | |
| fungi | so i compile foo.c in a job on change 1 and produce an artifact, i want to compile foo.c from change 2 and compare it against the result from change 1, but to do that i have to be sure that change 1's job has already completed to the point of producing that artifact | 18:13 |
| mordred | I do not believe the system is built with that use-case in mind. it's built to get the best speculative future artifact from the stack of dependent changes | 18:13 |
| mordred | hrm | 18:13 |
| clarkb | mordred: I think both providing and requires on a single job would address that | 18:13 |
| mordred | ok - that makes more sense to my brain | 18:13 |
| mordred | and yes - I think it's at least worth a try of providing and requiring there | 18:14 |
| mordred | my brain hurts now | 18:14 |
| fungi | i suppose the build for change 2 could start whenever but just busywait until it can find an artifact produced by the same job running for change 1 | 18:14 |
| fungi | that could make calculating job timeouts tough though | 18:15 |
| mordred | no - I think the simul provide/require is the trick | 18:15 |
| mordred | weird as it is for my brain | 18:16 |
| mordred | then the job content itself will still have to deal with "was there a job ahead of me that built an artifact that's in the list, grab it, else grab the published artifact" | 18:16 |
| clarkb | yup | 18:17 |
| mordred | although if it's docker images the base jobs will do that for them | 18:17 |
| fungi | so that would effectively create an implicit queue-specific semaphore for that job so that the builds for all changes ran that job one at a time in serial order | 18:17 |
| fungi | neat | 18:17 |
| mordred | (they'd still need to have custom content to pull the sha of the image in question at the start of the job before doig their build) | 18:17 |
| *** cdearborn has joined #zuul | 18:19 | |
| openstackgerrit | Monty Taylor proposed zuul/zuul-preview master: Check to make sure artifact has a url https://review.opendev.org/718186 | 18:25 |
| openstackgerrit | Monty Taylor proposed zuul/zuul-preview master: Check to make sure artifact has a url https://review.opendev.org/718186 | 18:28 |
| *** harrymichal has quit IRC | 18:37 | |
| *** harrymichal_ has joined #zuul | 18:37 | |
| *** harrymichal_ has quit IRC | 18:50 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: tox: allow tox to be upgraded https://review.opendev.org/690057 | 19:07 |
| *** y2kenny has joined #zuul | 19:14 | |
| *** saneax has quit IRC | 19:15 | |
| y2kenny | for job.files, is there a way to specify the project if multiple project uses the same job? | 19:17 |
| fungi | so one set of files if it's project a, different set of files for project b? i think you can stick files matchers in a job variant in the specific project-pipeline | 19:18 |
| y2kenny | ok | 19:18 |
| openstackgerrit | Merged zuul/zuul-preview master: Check to make sure artifact has a url https://review.opendev.org/718186 | 19:21 |
| *** bhavikdbavishi has quit IRC | 19:27 | |
| *** harrymichal has joined #zuul | 19:29 | |
| zbr | clarkb: if you can have a look at https://review.opendev.org/#/c/690057/ and tell me why it does not work on ubuntu it would be great, i am either too tired to duumb to see what is wrong. | 20:01 |
| zbr | apparently tox refuses to be let itself removed :p | 20:02 |
| clarkb | zbr: it looks like you remove it via python2 but its still installed via python3 | 20:03 |
| clarkb | note the deprecation warnings on the uninstall but the file path on the error | 20:04 |
| zbr | but the command -v lines are supposed to remove it from all pythons | 20:04 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Retry nodeenv creation https://review.opendev.org/717820 | 20:04 |
| zbr | unless some "pip" commands are really missing from path. maybe i should try using python -m to call pip instead. | 20:04 |
| zbr | i got the same impression that that command | xargs finds only pip, but my local tests worked fine. | 20:05 |
| fungi | yes, looking at the bionic job it's reporting a path of /usr/local/lib/python3.6/dist-packages/tox/__init__.py which means it was installed with pip for python3.6 (the default python on the platform) | 20:05 |
| zbr | on the same platforms. | 20:05 |
| zbr | fungi: in that case the package absent line should have already being removed it | 20:05 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: WIP: Enables whitelisting and configuring callbacks https://review.opendev.org/717260 | 20:06 |
| fungi | zbr: which package absent line? | 20:10 |
| fungi | this is not a distro package of tox, it's pip-installed tox using the distro's python3.6 interpreter (and some unknown installation of pip) | 20:10 |
| fungi | the "Remove tox package with pip(s) - root" task does: command -v pip pip2 pip3 | xargs -I{} sh -xc "{} uninstall -y tox || true" | 20:13 |
| fungi | which seems to only wind up running /usr/local/bin/pip | 20:14 |
| zbr | this means that the other pips are not that in path, i will try tomorrow to call pip as a module. | 20:18 |
| openstackgerrit | Albin Vass proposed zuul/zuul master: WIP: Enables whitelisting and configuring callbacks https://review.opendev.org/717260 | 20:22 |
| *** tjgresha has joined #zuul | 20:31 | |
| *** tjgresha has quit IRC | 20:31 | |
| *** hashar has joined #zuul | 20:38 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 20:38 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 20:38 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-pip: export ensure_pip_virtualenv_command https://review.opendev.org/718224 | 20:39 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] use ensure-pip in fetch-subunit-output test https://review.opendev.org/718225 | 20:39 |
| *** igordc has quit IRC | 20:47 | |
| *** zxiiro has quit IRC | 20:47 | |
| *** igordc has joined #zuul | 20:48 | |
| *** rlandy is now known as rlandy|biab | 21:10 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure pip role https://review.opendev.org/717639 | 21:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-pip: export ensure_pip_virtualenv_command https://review.opendev.org/718224 | 21:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] fetch-zuul-cloner: use ensure-pip https://review.opendev.org/717882 | 21:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] use ensure-pip in fetch-subunit-output test https://review.opendev.org/718225 | 21:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 21:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 21:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] fetch-zuul-cloner: use ensure-pip https://review.opendev.org/717882 | 21:45 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] use ensure-pip in fetch-subunit-output test https://review.opendev.org/718225 | 21:45 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: preinstall pip https://review.opendev.org/717663 | 21:45 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 21:45 |
| *** harrymichal has quit IRC | 21:52 | |
| *** rlandy|biab is now known as rlandy | 21:54 | |
| openstackgerrit | Merged zuul/nodepool master: Update tests for node-attributes https://review.opendev.org/714738 | 21:55 |
| openstackgerrit | Merged zuul/nodepool master: Support image uploads in 'info' CLI command https://review.opendev.org/712775 | 22:02 |
| openstackgerrit | Merged zuul/nodepool master: Fix shutdown ordering https://review.opendev.org/717134 | 22:02 |
| openstackgerrit | Merged zuul/nodepool master: Add GCE driver tests https://review.opendev.org/717133 | 22:02 |
| openstackgerrit | Merged zuul/zuul master: Annotate dynamic layout creation https://review.opendev.org/718160 | 22:13 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: [wip] ensure-tox: use ensure-pip role https://review.opendev.org/717663 | 22:35 |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: Update Fedora to 31 https://review.opendev.org/717657 | 22:35 |
| *** saneax has joined #zuul | 22:56 | |
| *** hashar has quit IRC | 22:56 | |
| jkt | wow, this is pretty strict, An unhandled exception occurred while running the lookup plugin 'env'. Error was a <class 'ansible.errors.AnsibleError'>, original message: Use of lookup modules that perform local actions on the executor is forbidden. | 23:11 |
| fungi | jkt: we've whitelisted them as folks have done sufficient analysis to make sure they can't open write access to arbitrary files or enable arbitrary code execution | 23:14 |
| jkt | yup, I'm on a pretty oldish Zuul | 23:14 |
| fungi | so it's possible that one's safe and just nobody's audited its implementation yet | 23:14 |
| jkt | besides, if it's really on the executor, then I should not use it anyway | 23:15 |
| fungi | right, that's the only place we blacklist them | 23:15 |
| jkt | a few more tries and I might have a working setup for coverage diffing | 23:15 |
| jkt | a few questions -- can I somehow return FAILURE instead of a FAILURE from a post_run playbook? | 23:16 |
| fungi | i don't understand the question | 23:16 |
| jkt | context is https://gerrit.cesnet.cz/c/ci/zuul-jobs-cesnet/+/2416/10 | 23:17 |
| fungi | i read that as you want a post_run playbook to return FAILURE instead of FAILURE, so i'm surely misunderstanding | 23:17 |
| jkt | doh | 23:17 |
| jkt | I would like a post_run playbook to return FAILURE instead of POST_FAILURE that it can return now | 23:17 |
| fungi | ahh, interesting, i'm not aware of a way, no... usually the idea is that you put things you want to represent actual failure conditions when they don't work into the run phase | 23:18 |
| fungi | so POST_FAILURE is how zuul indicates that a failure condition was found in a post phase playbook | 23:19 |
| fungi | as opposed to during the run phase | 23:19 |
| fungi | and then during the pre phase, errors are retried until RETRY_LIMIT condition is returned | 23:19 |
| fungi | so the phases are mostly about how the system behaves when a playbook doesn't work | 23:20 |
| jkt | I can do that, but it means that I cannot "just" inherit from my "main build job" in the "build prev version + compute coverage diff" job definition | 23:21 |
| jkt | no worries, it's just very slightly inconvenient | 23:21 |
| jkt | second Q, can I somehow set job.success_url as a relative suffix from my job? | 23:22 |
| jkt | I would like to keep on whatever setup is already in place for uplaoding logs and setting that absolute URL in log_url | 23:22 |
| jkt | and then, if my job made it as far as to the very end, which means that there's that coverage-diff.html in my job dir, only then append a relative suffix to the URL | 23:22 |
| jkt | if not, as the patch above is doing, then the user might click to a link which doesn't work, but once again, this should be relatively rare | 23:23 |
| clarkb | you might want ot add it as an artifact instead? | 23:23 |
| clarkb | then you get a nice link to it in the dashboard | 23:23 |
| fungi | yeah, success_url and friends are somewhat inflexible in that you only get one | 23:24 |
| jkt | clarkb: I can do that, but given that this job cannot really produce anything extra apart from this HTML, I would like to save that extra click | 23:24 |
| fungi | so if you set it to anything special then it makes it hard to get to the build result details | 23:24 |
| fungi | well, i mean, the job likely also produces a console log, job metadata, et cetera | 23:25 |
| fungi | so if the job fails, getting to that information to debug becomes a fair bit harder | 23:25 |
| jkt | yeah | 23:25 |
| *** tosky has quit IRC | 23:26 | |
| clarkb | I'm not sure how to set up relative urls fwiw | 23:27 |
| clarkb | but I do recall we switched away from that mechanism entirely once the dasboard could show builds | 23:27 |
| fungi | though i suppose if it's just success_url then maybe that's less of a concern as long as the job doesn't do something it's not supposed to and succeed anyway | 23:27 |
| jkt | yeah, mine instance of Zuul is on 3.8, no fancy dashboards there | 23:28 |
| fungi | there is some basic templating for the url parameters, refreshing my recollection from docs | 23:28 |
| fungi | hrm, we actually no longer document it at all that i can find | 23:30 |
| clarkb | fungi: ya I think the dashboard updates effectively broke the old system | 23:31 |
| clarkb | so it was an either or | 23:31 |
| fungi | well, we don't seem to have documented it at all for v3 | 23:32 |
| fungi | looking back through git history | 23:32 |
| fungi | it looks like it's still implemented in QueueItem.formatProvisionalJobResult() and there's one unit test of it | 23:33 |
| fungi | but yeah, basically an undocumented feature and seems to have been so at least since v3 | 23:34 |
| fungi | it's definitely not something we generally encourage relying on because of the reasons mentioned above. it looks like it can template in anything formatUrlPattern() supports | 23:36 |
| fungi | which is basically the set of: change, pipeline, tenant, buildset, job, build | 23:36 |
| jkt | nice, but I think that won't help me | 23:37 |
| fungi | everything else is statically defined in the value set in the config, so at best also subject to yaml's templating features, which are limited at best | 23:38 |
| fungi | and interpreted at configuration load time | 23:38 |
| jkt | btw, funny intermezzo #123, this time related to our Swift configuration -- it always sends out "content-encoding: deflate", so Ansible's get_url won't retrieve it properly, and I think I have to call out to `curl --compress` | 23:39 |
| fungi | we had to take extra care to only upload compressed files if we want them downloaded in a compressed format (so stuff like .tar.gz files), and otherwise upload uncompressed and rely on swift's built-in compression for serving up things we want retried plain | 23:49 |
| fungi | s/retried/retrieved/ | 23:49 |
| jkt | fungi: well, I'm unsure if it's a bug on my side or somewhere else, but `wget https://object-store.cloud.muni.cz/swift/v1/ci-logs-CzechLight-internal/14/2414/19/check/f29-gcc/eb5f0f5/coverage.xml` doesn't return an actual XML for me | 23:55 |
| clarkb | jkt: are you using the zuul-jobs swift upload role? | 23:56 |
| clarkb | basically it was uploading a compressed verson of a file then also setting the encoding type to be compressed. At the time I THink the idea was that everything was compressed and the values agreed with each other. Problem is you run into that behavioryou see | 23:57 |
| clarkb | instead what we can do is upload compressed (and it is stored compressed) but set the encoding type to uncompressed and swift seems to figure it out at the web server level | 23:58 |
| clarkb | let me see ifI can find the change | 23:58 |
| jkt | clarkb: yes; mine is at https://gerrit.cesnet.cz/plugins/gitiles/ci/zuul-jobs/+/2da7fecff5d1650b8767a7d294deeff3e6fc54bd%5E%21/ which I hope is a cherry-pick of mine c31cd37164ae3e24c70a9de078d8f637386ce8ec upstream | 23:58 |
| clarkb | 0e23325a12651756faf5b99dda308a93788bd9db is the commit that fixes it | 23:59 |
| clarkb | though there were a couple followups after that the cleaned up some edge case stuff | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!