| *** sgw has joined #zuul | 01:55 | |
| *** bhavikdbavishi has joined #zuul | 02:22 | |
| *** bhavikdbavishi1 has joined #zuul | 03:04 | |
| *** bhavikdbavishi has quit IRC | 03:05 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 03:05 | |
| *** bhavikdbavishi has quit IRC | 03:38 | |
| *** bhavikdbavishi has joined #zuul | 04:15 | |
| tobiash | corvus: I left a comment on 670335 | 04:26 |
|---|---|---|
| tobiash | corvus: I left a question on 670666 | 04:31 |
| *** bolg has joined #zuul | 04:43 | |
| *** raukadah is now known as chandankumar | 05:26 | |
| *** badboy has joined #zuul | 05:32 | |
| *** AJaeger has quit IRC | 05:49 | |
| *** AJaeger has joined #zuul | 05:53 | |
| *** saneax has quit IRC | 06:14 | |
| *** saneax has joined #zuul | 06:26 | |
| *** zbr has quit IRC | 06:50 | |
| *** fungi has quit IRC | 06:50 | |
| *** odyssey4me has quit IRC | 06:50 | |
| *** daniel2 has quit IRC | 06:50 | |
| *** persia has quit IRC | 06:50 | |
| *** mugsie has quit IRC | 06:50 | |
| *** AJaeger has quit IRC | 06:50 | |
| *** jhesketh has quit IRC | 06:50 | |
| *** aluria has joined #zuul | 07:04 | |
| *** AJaeger has joined #zuul | 07:16 | |
| *** jhesketh has joined #zuul | 07:16 | |
| *** tosky has joined #zuul | 07:16 | |
| *** zbr has joined #zuul | 07:16 | |
| *** fungi has joined #zuul | 07:16 | |
| *** odyssey4me has joined #zuul | 07:16 | |
| *** daniel2 has joined #zuul | 07:16 | |
| *** persia has joined #zuul | 07:16 | |
| *** mugsie has joined #zuul | 07:16 | |
| *** openstackgerrit has quit IRC | 07:18 | |
| *** panda has quit IRC | 07:19 | |
| *** panda has joined #zuul | 07:21 | |
| badboy | what is the logic behind creating directories by post-logs.yaml? | 07:32 |
| badboy | I am wondering about the first part of: ./62/<GERRIT_REVIEW>/<GERRIT_PATCHSET>/<JOB_NAME>/<JOB_NAME>/d3409aa/job-output.json.gz | 07:34 |
| AJaeger | badboy: It's not /62/ but /<last-two-digits-of-gerrit-review>/<gerrit-review> | 07:38 |
| AJaeger | So, 62 are only the last digit, review 670335 will use 35. Storing in 100 directories instead of a single one is better for navigating filesystems | 07:38 |
| *** hashar has joined #zuul | 07:47 | |
| *** saneax has quit IRC | 07:50 | |
| *** jangutter has joined #zuul | 07:55 | |
| tobiash | infra-root: it looks like we lost the openstackgerrit bot here | 08:14 |
| badboy | AJaeger: thank you. Is there a way to display all logs in the web ui? | 08:16 |
| *** AJaeger has quit IRC | 08:26 | |
| *** jhesketh has quit IRC | 08:26 | |
| *** zbr has quit IRC | 08:27 | |
| *** fungi has quit IRC | 08:27 | |
| *** odyssey4me has quit IRC | 08:27 | |
| *** daniel2 has quit IRC | 08:27 | |
| *** persia has quit IRC | 08:27 | |
| *** mugsie has quit IRC | 08:27 | |
| *** panda has quit IRC | 08:29 | |
| *** panda has joined #zuul | 08:31 | |
| *** persia has joined #zuul | 08:35 | |
| *** daniel2 has joined #zuul | 08:35 | |
| *** zbr has joined #zuul | 08:36 | |
| *** odyssey4me has joined #zuul | 08:36 | |
| *** AJaeger has joined #zuul | 08:36 | |
| *** mugsie has joined #zuul | 08:37 | |
| *** fungi has joined #zuul | 08:37 | |
| *** jhesketh has joined #zuul | 08:37 | |
| migi | Hi, can a job.provides and job.requires be within one project-template definition ? I have two jobs that are in the same project-template for the same check and they seem to not honor proper order of starting those jobs | 08:52 |
| *** saneax has joined #zuul | 08:54 | |
| * migi found the asnwer it's only for Depepends-On | 08:56 | |
| tobiash | migi: you want to use job dependencies for ordering within a buildset | 09:02 |
| migi | tobiash: yep | 09:03 |
| migi | tobiash: I have 3 jobs first is building artefact and 2 are using it for the same changeset | 09:03 |
| tobiash | migi: you can use zuul_return for transferring information to the next jobs in the buildset | 09:04 |
| migi | tobiash: and found that job.requires and job.provides are not scheduling jobs in the proper order if that's same changeset | 09:04 |
| migi | tobiash: ok will look into that | 09:04 |
| migi | tobiash: and if I use job.requires ad job.provides then use job.dependencies between those it should work, right? | 09:05 |
| tobiash | migi: yes, the information should be there | 09:05 |
| migi | tobiash: so proper scheduling will be by job.dependencies and job.requires/provides will provide info | 09:05 |
| migi | ok, will check that | 09:05 |
| tobiash | ++ | 09:05 |
| migi | tobiash: thanks ! | 09:05 |
| tobiash | no problem | 09:06 |
| *** sshnaidm is now known as sshnaidm|rover | 09:29 | |
| *** snapiri has quit IRC | 09:31 | |
| *** johnsom has quit IRC | 10:06 | |
| *** johnsom has joined #zuul | 10:07 | |
| *** pcaruana has joined #zuul | 11:33 | |
| *** snapiri has joined #zuul | 11:42 | |
| *** weshay|rover is now known as weshay | 11:51 | |
| zbr | https://review.opendev.org/#/c/670044/ -- small improvement on log collection | 12:33 |
| *** electrofelix has joined #zuul | 12:36 | |
| *** electrofelix has quit IRC | 12:56 | |
| *** jeliu_ has joined #zuul | 12:58 | |
| *** openstackgerrit has joined #zuul | 13:28 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: WIP: Route streams to different zones via finger gateway https://review.opendev.org/664965 | 13:28 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Support ssl encrypted fingergw https://review.opendev.org/664950 | 13:28 |
| *** electrofelix has joined #zuul | 13:33 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: WIP: Route streams to different zones via finger gateway https://review.opendev.org/664965 | 13:37 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Support ssl encrypted fingergw https://review.opendev.org/664950 | 13:37 |
| *** michael-beaver has joined #zuul | 13:52 | |
| *** gtema_ has joined #zuul | 13:58 | |
| AJaeger | corvus: could you review https://review.opendev.org/667813 again, please? The 2 weeks waiting period is over... | 14:04 |
| *** gtema_ has quit IRC | 14:06 | |
| *** gtema_ has joined #zuul | 14:06 | |
| fungi | AJaeger: when you get a moment, do you think the opensuse bit in https://review.opendev.org/670736 is correct, or should it be adjusted while i'm in there? | 14:35 |
| AJaeger | fungi: there's also a git-review rpm package but that is version 1.25, so what you have looks fine - thanks | 14:38 |
| corvus | tobiash: thanks, replied on both of those | 14:38 |
| corvus | AJaeger: +2 -- would you mind sending a followup announcement for that one to let people know it's done? (since they might want to update a local copy and/or remove the old config) | 14:40 |
| corvus | AJaeger: i can coordinate +W and releasing the announcement | 14:40 |
| fungi | AJaeger: cool, was mostly wondering whether the python3 stuff was there automatically and if pip was included by default or whether any additional steps would be required to get it. thanks! | 14:41 |
| *** irclogbot_0 has quit IRC | 14:41 | |
| corvus | fungi: fedora doesn't need epel-release | 14:42 |
| tobiash | corvus: thanks for the explanation, I revised my votes | 14:43 |
| fungi | corvus: got it, so we need to split the fedora and centos/rhel sections i guess | 14:43 |
| corvus | yeah. or we could drop centos support. this was supposed to be a one-liner. | 14:45 |
| fungi | i'm good either way, just we currently list platforms which require some deep knowledge of python packaging (and most of them have their own quirk or two) | 14:47 |
| fungi | it got more complicated when we switched to a new enough gerrit that the distro packaged version of git-review on most platforms is not new enough to support | 14:48 |
| AJaeger | fungi: it looked fine but my system is not vanilla ;( | 14:49 |
| AJaeger | corvus: followup to announce mailing list or to normal zuul one? | 14:49 |
| corvus | AJaeger: announce | 14:49 |
| clarkb | as an alternative could add another container to the setup for using git and git-review. Then we can preinstall those tools there. (sort of clunky if you want it to mimic development process though) | 14:50 |
| clarkb | (and is likely to trigger more questions around "my volume doesn't work" like we've had recently) | 14:51 |
| *** irclogbot_3 has joined #zuul | 14:52 | |
| fungi | debian-buster and ubuntu-disco have new enough git-review to work with latest gerrit, but that's about it looks like | 14:52 |
| fungi | maybe fedora too? i don't know how to best look up package versions in it | 14:53 |
| fungi | nor tumbleweed | 14:53 |
| AJaeger | corvus: sent | 14:53 |
| clarkb | fedora and suse both have package search tools: https://apps.fedoraproject.org/packages/ and https://software.opensuse.org/find | 14:54 |
| fungi | i figured they did, i was just taking longer than i had hoped to hunt them down via web searches | 14:54 |
| clarkb | git-review is 1.28 on tumbleweed | 14:54 |
| clarkb | 1.28 on fedora 30 but 29 is 1.26 | 14:54 |
| AJaeger | clarkb: openSUSE Leap 15.1 has 1.25 | 14:55 |
| fungi | so maybe we could limit the quickstart instructions to debian-buster, ubuntu-disco, opensuse-tumbleweed and fedora-30 | 14:55 |
| fungi | challenge is we test the quickstart routine on ubuntu-bionic | 14:56 |
| *** zer0c00l has quit IRC | 14:56 | |
| corvus | maybe let's just split centos out for now, and when centos 8 is out, we revisit? | 14:56 |
| *** zer0c00l has joined #zuul | 14:56 | |
| fungi | that sounds fine to me | 14:57 |
| fungi | will update | 14:57 |
| fungi | does fedora need pip installed explicitly, or is it there by default? | 14:58 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul master: Correct git-review install in CentOS 7 quick-start https://review.opendev.org/670736 | 14:59 |
| fungi | new patchset ^ is assuming it's already there | 14:59 |
| corvus | clarkb: can you take a look at https://review.opendev.org/670612 ? http://logs.openstack.org/12/670612/1/gate/nodepool-functional-openstack/d27ee42/nodepool/nodepool.log | 15:00 |
| corvus | clarkb: is it possible the glean work broke that? | 15:01 |
| clarkb | corvus: I think that failed because we needed a dib release. A recheck should work now that ianw did that release | 15:01 |
| clarkb | corvus: the src jobs all worked because they don't rely on a release but the non src job needed the release which is done now | 15:01 |
| corvus | so dib and glean need non-src jobs added | 15:02 |
| clarkb | I would say at least the one ya | 15:02 |
| clarkb | as a sanity check | 15:02 |
| corvus | clarkb: i'll propose those patches, thx | 15:02 |
| *** chandankumar is now known as raukadah | 15:14 | |
| *** hashar has quit IRC | 15:25 | |
| *** igordc has joined #zuul | 15:27 | |
| openstackgerrit | Graham Hayes proposed zuul/nodepool master: Implement an Azure driver https://review.opendev.org/554432 | 15:32 |
| openstackgerrit | Merged zuul/zuul master: Build layout of non-live items with config updates https://review.opendev.org/670335 | 15:33 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Add "supercedes" pipeline option https://review.opendev.org/670670 | 15:34 |
| *** gtema_ has quit IRC | 15:36 | |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: [WIP] Verify Operator Pod Running https://review.opendev.org/670395 | 15:41 |
| AJaeger | corvus: zuul is not putting https://review.opendev.org/#/c/667813/ (zuul/zuul-base-jobs) into the gate queue. I do not see jobs configured for it, should they be in zuul/project-config? | 15:44 |
| corvus | AJaeger: yes, or we can move them in-repo in zuul-tests.d/ | 15:45 |
| corvus | AJaeger: do you have time to do that, or should i? | 15:45 |
| AJaeger | corvus: I would prefer if you can do it... Will have later time to review. | 15:46 |
| AJaeger | thanks, corvus | 15:46 |
| corvus | k | 15:46 |
| openstackgerrit | James E. Blair proposed zuul/zuul-base-jobs master: Add linter/docs jobs https://review.opendev.org/670872 | 15:52 |
| corvus | AJaeger: this may take a couple of round-trips | 15:53 |
| AJaeger | corvus: fun ;) | 15:53 |
| corvus | can i get 2x+2 on that now? i'll force-merge it then (that's going to be the easiest way to bootstrap this) | 15:54 |
| *** jamesmcarthur has joined #zuul | 15:54 | |
| corvus | AJaeger, clarkb, fungi, tobiash: ^ | 15:55 |
| AJaeger | corvus: left a comment | 15:55 |
| AJaeger | corvus: read your comment - you have two +2s ;) | 15:56 |
| openstackgerrit | Merged zuul/zuul-base-jobs master: Add linter/docs jobs https://review.opendev.org/670872 | 15:57 |
| sgw | Morning all, I have been working on a Zuul job to interface with openSUSE OBS for StarlingX, I have a very early review (https://review.opendev.org/#/c/670363/) realize this is my first Zuul job and crack at ansible. I would appreciate comments, but am looking for some specific help that is not clear in the docs. I want to run on an openSUSE-15 node to start with to ensure I have the correct "osc" command, when I tried with Ubuntu node, | 15:58 |
| sgw | I could install osc, but it failed to run correctly. Thanks in advance. | 15:58 |
| openstackgerrit | James E. Blair proposed zuul/zuul-base-jobs master: WIP: test https://review.opendev.org/670874 | 15:58 |
| sgw | This might be useful for the openstack/rpm-packaging repo once I stabilize this version (I hope) | 15:59 |
| corvus | sgw: the 'nodeset' job attribute tells zuul what kind of node to use | 16:00 |
| corvus | sgw: do you see the error message from zuul on that change? | 16:01 |
| AJaeger | sgw: see my -1 for the fix | 16:01 |
| AJaeger | sgw: did you see http://logs.openstack.org/63/670363/15/check/flock-opensuse-build/3683a4d/job-output.txt.gz#_2019-07-13_01_25_42_896749 | 16:02 |
| AJaeger | Why are you adding an Ubuntu repo on an openSUSE node? | 16:02 |
| AJaeger | sgw: and looking at http://logs.openstack.org/63/670363/15/check/flock-opensuse-build/3683a4d/job-output.txt.gz#_2019-07-13_01_25_50_277500 you seem to have a quoting problem | 16:03 |
| *** hashar has joined #zuul | 16:05 | |
| *** hwangbo has joined #zuul | 16:05 | |
| sgw | AJaeger: ah typo in name! Not sure about adding the Ubuntu repo in openSUSE, the opensuse version points to the pre-opensuse role that should not have Ubuntu repos | 16:07 |
| sgw | AJaeger: Ultimately I want to make this usable across multiple StarlingX repos (I guess a starlingx-base job), so I should name it starlingx-osc-build? | 16:07 |
| corvus | sgw: if you didn't see the error message zuul left for you, try pressing the "toggle ci" button in the bottom left | 16:10 |
| sgw | AJaeger: regarding the quoting, I use "rdelete -r -m 'rpm-packing CI Cleanup' {{ test_project }}" in the ansible code, is that not correct for nesting quotes, the api failure below only uses one set of quotes. | 16:10 |
| openstackgerrit | James E. Blair proposed zuul/zuul-base-jobs master: Add gate jobs https://review.opendev.org/670874 | 16:12 |
| corvus | clarkb, AJaeger: ^ | 16:12 |
| AJaeger | sgw: starlingx-osc-build looks fine. For the rest: Try it out ;) You now know some of the failures to look for ;) | 16:13 |
| openstackgerrit | Merged zuul/zuul-base-jobs master: Add gate jobs https://review.opendev.org/670874 | 16:14 |
| *** hwangbo has quit IRC | 16:17 | |
| openstackgerrit | Andreas Jaeger proposed zuul/project-config master: Update promote pipeline precedence https://review.opendev.org/670879 | 16:22 |
| openstackgerrit | Merged zuul/zuul-base-jobs master: Use opendev.org as git location https://review.opendev.org/667813 | 16:26 |
| openstackgerrit | Merged zuul/zuul master: Logging of reconfiguration durations in Zuul https://review.opendev.org/670773 | 16:28 |
| AJaeger | corvus: 670879 was done for openstack/project-config but forgotten to update here | 16:29 |
| corvus | AJaeger: thx | 16:31 |
| openstackgerrit | Merged zuul/project-config master: Update promote pipeline precedence https://review.opendev.org/670879 | 16:32 |
| openstackgerrit | Jeff Liu proposed zuul/zuul-operator master: [WIP] Verify Operator Pod Running https://review.opendev.org/670395 | 16:33 |
| openstackgerrit | Jeremy Stanley proposed zuul/zuul master: Correct git-review install in CentOS 7 quick-start https://review.opendev.org/670736 | 16:38 |
| AJaeger | corvus: promote job works - https://zuul-ci.org/docs/zuul-base-jobs/ was updated | 16:39 |
| *** mattw4 has joined #zuul | 16:42 | |
| openstackgerrit | Merged zuul/zuul master: Annotate canMerge check with event id https://review.opendev.org/670494 | 16:42 |
| *** hashar has quit IRC | 16:59 | |
| *** AJaeger_ has joined #zuul | 17:00 | |
| *** mattw4 has quit IRC | 17:02 | |
| *** AJaeger has quit IRC | 17:03 | |
| corvus | i'm looking into the abort_cleanup test failure that's been creeping up, and i think it's showing us a bug | 17:06 |
| *** igordc has quit IRC | 17:07 | |
| fungi | i saw it yesterday on a doc patch, but it wasn't entirely obvious what's wrong with the test | 17:11 |
| *** tosky has quit IRC | 17:13 | |
| *** tosky has joined #zuul | 17:13 | |
| *** bolg has quit IRC | 17:19 | |
| *** mattw4 has joined #zuul | 17:19 | |
| *** jeliu_ has quit IRC | 17:23 | |
| *** altlogbot_0 has quit IRC | 17:25 | |
| *** AJaeger_ is now known as AJaeger | 17:25 | |
| *** irclogbot_3 has quit IRC | 17:26 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Run cleanup playbooks in job thread https://review.opendev.org/670888 | 17:27 |
| corvus | fungi, tobiash, tristanC: ^ | 17:27 |
| corvus | oh i see more code i can delete | 17:27 |
| *** irclogbot_0 has joined #zuul | 17:30 | |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Run cleanup playbooks in job thread https://review.opendev.org/670888 | 17:31 |
| *** irclogbot_0 has quit IRC | 17:32 | |
| fungi | cleanup the cleanup | 17:32 |
| corvus | fungi, tobiash, tristanC: ^ there i think that's got it | 17:32 |
| fungi | what's a "stob job"? an actual thing, or a typo? | 17:32 |
| corvus | fungi: a really consistent typo | 17:33 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Run cleanup playbooks in job thread https://review.opendev.org/670888 | 17:33 |
| fungi | aha! your p was upside down | 17:34 |
| corvus | well, that's a q | 17:34 |
| corvus | (which is why you mind your p's and q's) | 17:34 |
| fungi | now it makes a lot more sense, thanks! | 17:34 |
| SpamapS | did we remove add-build-sshkey from zuul-jobs recently? | 17:35 |
| * SpamapS is failing all jobs at the moment | 17:35 | |
| corvus | SpamapS: we did this: http://lists.zuul-ci.org/pipermail/zuul-announce/2019-July/000045.html | 17:36 |
| SpamapS | OH! I remember that. k... | 17:36 |
| * SpamapS really needs to make a gated mirror | 17:37 | |
| corvus | i'm hoping we don't change that too often | 17:38 |
| SpamapS | It would also be really cool if we could add something that lets people carry the new one before the migration happens. I remember I made the change the day that announcement was made, and it broke everything.. so then.. I forgot. :-/ | 17:39 |
| SpamapS | Yeah I know it's a rare thing. | 17:39 |
| fungi | the announcement suggested just adding the connection in advance should be safe? | 17:39 |
| SpamapS | so not like, "do this now!" just.. would be nice to be able to express "this host is an alias for that host" | 17:39 |
| corvus | SpamapS: oh, i think you can have both connections | 17:39 |
| SpamapS | like the way shadows works | 17:40 |
| *** irclogbot_3 has joined #zuul | 17:40 | |
| SpamapS | corvus: having both connections isn't something I remember reading, so I just missed the whole thing | 17:40 |
| sgw | AJaeger: question, I tested most of the ansible code locally using the quickstart zuul setup, particularly the "osc" related commands and they all work locally, but are failing when run via the CI, any thoughts on what would be different? | 17:40 |
| SpamapS | as usual, I'm the canary and as such I was dead when the better advice was given. ;) | 17:41 |
| *** altlogbot_3 has joined #zuul | 17:41 | |
| corvus | SpamapS: yeah, i think we broke you, reverted, then came up with better advice, probably while you were cleaning off your firefighting gear | 17:42 |
| SpamapS | 'sok. Price I pay to ride the lightning. ;) | 17:42 |
| SpamapS | FYI, that original message is not very clear, the connection change should be from https://git.zuul-ci.org/ -> https://opendev.org/zuul/ (one can infer this from the linked patch, but it's not 100% clear) | 17:47 |
| corvus | good point; AJaeger, should we send out a followup? | 17:48 |
| openstackgerrit | Merged zuul/zuul master: Handle existing broken config in job updates https://review.opendev.org/670666 | 17:49 |
| AJaeger | sgw: no idea, haven't used osc for some time. cmurphy or dirk might be able to help, they should be on #openstack-infra | 17:54 |
| AJaeger | sgw: the only idea I had was that quoting is wrong - looking at hte error message, so I would have removed the outer "" | 17:54 |
| AJaeger | corvus: I can do that... | 17:54 |
| sgw | AJaeger: I can try that, it worked locally with the quickstart guide, so I thought it should be the same | 17:59 |
| *** armstrongs has joined #zuul | 18:01 | |
| armstrongs | hey guys im trying to use role_branch: "{{ zuul.branch }}" passed as extra_vars but it isn't substituting can it not be done this way? | 18:03 |
| armstrongs | this is in my .zuul.yaml | 18:04 |
| *** saneax has quit IRC | 18:04 | |
| *** igordc has joined #zuul | 18:05 | |
| AJaeger | armstrongs: could you give a link, please? | 18:05 |
| fungi | certain sorts of pipeline triggers won't have a branch | 18:06 |
| fungi | it depends on whether the event details include one | 18:06 |
| armstrongs | this is a github pull request trigger | 18:06 |
| armstrongs | http://paste.openstack.org/show/754404/ | 18:07 |
| fungi | presumably that will have a branch then, so that's probably not the cause | 18:08 |
| armstrongs | and my file looks like | 18:09 |
| armstrongs | http://paste.openstack.org/show/754405/ | 18:09 |
| corvus | armstrongs: maybe extra-vars can't reference inventory variables? | 18:11 |
| corvus | hrm, that seems to work with a local test of ansible | 18:14 |
| armstrongs | yeah im a bit confused | 18:14 |
| armstrongs | isnt the quotes or anything is it | 18:14 |
| corvus | quotes look good | 18:14 |
| armstrongs | ok i think it may be something to do with the var being passed, so its already a token in a jinja file, so its actually doing the correct thing and tokenizing it with the value. It just so happens the variable is a variable if that makes sense. | 18:19 |
| corvus | armstrongs: are you using it an an ansible "template" task? | 18:20 |
| armstrongs | yes | 18:20 |
| armstrongs | so a token to a token i guess doesnt work | 18:20 |
| corvus | i think it should... | 18:21 |
| armstrongs | so did I :) | 18:21 |
| armstrongs | if i pass it not as a variable and hardcoded its fine | 18:21 |
| openstackgerrit | Merged zuul/nodepool master: Add functional jobs to gate https://review.opendev.org/670612 | 18:21 |
| armstrongs | but with token i get version: '{{ zuul.branch }}' templated out | 18:22 |
| armstrongs | rather than the branch | 18:22 |
| corvus | armstrongs: this works for me: http://paste.openstack.org/show/754407/ | 18:23 |
| armstrongs | yeah which is exactly what mine should do | 18:24 |
| armstrongs | damn | 18:24 |
| corvus | armstrongs: here's a simple check in zuul: https://review.opendev.org/670897 http://logs.openstack.org/97/670897/1/check/sandbox-test/869a147/job-output.txt.gz#_2019-07-15_18_26_29_579198 | 18:27 |
| corvus | armstrongs: that doesn't have the template file, but i'm pretty sure if that works, it should be behaving the same in zuul | 18:28 |
| armstrongs | yeah thanks for the help, will debug this some more. | 18:30 |
| *** electrofelix has quit IRC | 18:31 | |
| armstrongs | at least now i know im doing something that should work and wasnt missing something | 18:31 |
| corvus | ++ | 18:32 |
| SpamapS | hrm.. so.. I switched the connection.. but now I'm not even getting any zuul-jobs roles | 18:35 |
| SpamapS | Like, zuul-jobs isn't even in the included roles section | 18:35 |
| SpamapS | (basing this on the error message I'm seeing) | 18:36 |
| *** hwangbo has joined #zuul | 18:36 | |
| SpamapS | looks like we were using the short project name | 18:38 |
| SpamapS | and now that doesn't work | 18:38 |
| *** panda has quit IRC | 18:38 | |
| *** panda has joined #zuul | 18:40 | |
| SpamapS | nope | 18:40 |
| SpamapS | explicit name doesn't fix it | 18:41 |
| SpamapS | gah | 18:44 |
| SpamapS | so the name from the projects UI does not map to the name needed in the roles statement | 18:44 |
| SpamapS | *THAT* is really frustrating | 18:44 |
| SpamapS | Maybe let's *never* move things. Ever. | 18:44 |
| SpamapS | FYI, the projects UI says that zuul-jobs name is 'opendev.org/zuul-jobs'. But zuul-base-jobs says 'opendev.org/zuul/zuul-jobs'. My connection info is https://opendev.org/zuul/.. but I'm guessing needs to drop the /zuul/ | 18:47 |
| SpamapS | all of this to say.. this sucks.. we're down.. I'm stressed.. would have appreciated a more thorough guide. | 18:48 |
| SpamapS | Also, at some point, I'd love to hear from all of you what I should do differently so that these things don't *always* break me. | 18:52 |
| SpamapS | The magic incantation btw, was to set the connection to opendev.org , but then make all project names fully explicit including with the zuul namespace | 18:57 |
| SpamapS | Only had to restart zuul 4 times and push 2 wrong commits to project config before it worked :-P | 18:58 |
| *** hashar has joined #zuul | 18:58 | |
| AJaeger | SpamapS: sorry to hear this ;( | 18:59 |
| AJaeger | Here's one change I just did, hope that's correct https://review.opendev.org/670895 | 19:00 |
| *** armstrongs has quit IRC | 19:06 | |
| corvus | AJaeger: that's a link to the inaugust.com website source? | 19:07 |
| clarkb | for consistency's sake ( and thus hopefully easier consumption) maybe connections should always be a hostname and not a url? | 19:12 |
| clarkb | I'm sure that breaks assumptions elsewhere, but then we'd always be forced to fullyqualify names in these instances | 19:12 |
| corvus | clarkb: can you clarify? connections are generally identified by canonical hostnames.... | 19:14 |
| corvus | SpamapS: maybe zuul-base-jobs isn't ready for production use yet? maybe just use it as a template and run with a fork of it. it doesn't change much. | 19:14 |
| corvus | all the roles are in zuul-jobs anyway | 19:15 |
| AJaeger | corvus: yes, to document src/zuulv3/getting-started.rst on the website which has an example | 19:16 |
| clarkb | corvus: it sounds like part of spamaps problem was specifying the connection as https://opendev.org/zuul instead of https://opendev.org so that the full path for the project name didn't line up? | 19:16 |
| *** bhavikdbavishi has quit IRC | 19:16 | |
| corvus | AJaeger: i just don't understand why we're talking about mordred's personal website. that seems like it should be way out of scope here. | 19:16 |
| clarkb | corvus: his project was logically opendev.org/zuul-jobs but the one in the job was opendev.org/zuul/zuul-jobs | 19:16 |
| corvus | AJaeger: i don't think we collectively maintain that | 19:16 |
| clarkb | I'm suggesting that if we take away the freedom for that flexibility then we'll all be forced to use fully qualified names in jobs | 19:17 |
| corvus | clarkb: i don't understand the suggestion | 19:17 |
| corvus | clarkb: spell it out for me please :) | 19:17 |
| AJaeger | corvus: I searched in codesearch for an example of the change that needed to be done and that was the only one that came up - and wanted to show ti to explain on what I think is the change that needs to be done for the connection | 19:18 |
| clarkb | corvus: stop allowing conenctions to specify a prefix like /zuul. Then the project names will always match in the jobs. So https://opendev.org/zuul in the connection would be invalid | 19:18 |
| corvus | clarkb: that is currently not possible for a gerrit or github connection; i don't think we could make that assertion for a git connection. | 19:19 |
| clarkb | right I'm sure it breaks things in the driver itself (since a gerrit might actually be hosted at some prefix) | 19:20 |
| clarkb | but if we removed flexibility to foot gun yourself there it might help people in SpamapS' situation | 19:20 |
| clarkb | (not sure how to do that ) | 19:20 |
| corvus | clarkb: i'm saying that you already can't drop a prefix with gerrit | 19:20 |
| corvus | you only could conceivably do that with git (and that's because git repos are arbitrary urls) so we just can't force the issue | 19:21 |
| corvus | i think the underlying issue is that it may not have been possible to make this change in a non-breaking manner, so our earlier email that suggested it was possible was just wrong | 19:21 |
| clarkb | ya I think there may be a couple problems here. One being you can't do it without breaking, the other being we apparently have enough flexibility in the config that even when you do update it and restart its still wrong | 19:22 |
| clarkb | (and maybe docs are the way forward there given the vast variety of content on the internet) | 19:23 |
| corvus | we could have tested various scenarios. but at the end of the day, we're volunteer reviewers, and i don't know that we can catch everything. | 19:24 |
| clarkb | and docs for this should say something like an http path prefix should be the root of where git repos are hosted. Though saying that "out loud" I'm not sure that helps clarify things | 19:25 |
| corvus | clarkb: we renamed the repo. there's no way around that. | 19:25 |
| clarkb | corvus: yes I know. But SpamapS additionally had trouble with the path prefix if I'm reading his messages correctly | 19:26 |
| clarkb | so part of the problem was not knowing what the new config should be | 19:26 |
| corvus | clarkb: i'm saying the path prefix is part of the name | 19:26 |
| corvus | we could have said "roles: zuul-jobs" and it would have worked everywhere, but we said "roles: opendev.org/zuul/zuul-jobs". we made it a fully qualified name, and then we changed that name. | 19:27 |
| corvus | hrm, wait... | 19:30 |
| corvus | maybe it would have worked if SpamapS had just not added 'zuul/' to the connection? | 19:31 |
| *** hashar has quit IRC | 19:31 | |
| clarkb | yes that is my assumption | 19:31 |
| corvus | well, considering we didn't actually tell people how to set up the connection, i guess it's hard to fault | 19:33 |
| corvus | we should have pointed people to https://opendev.org/zuul/zuul/src/branch/master/doc/source/admin/examples/etc_zuul/zuul.conf#L23-L26 | 19:33 |
| corvus | yeah, i'm pretty sure this could have been done seamlessly had we provided better instructions. a review comment or mailing list reply suggesting that we have better instructions would have been welcome. | 19:35 |
| *** jamesmcarthur has quit IRC | 19:51 | |
| *** altlogbot_3 has quit IRC | 20:10 | |
| *** jeliu_ has joined #zuul | 20:12 | |
| *** altlogbot_3 has joined #zuul | 20:14 | |
| fungi | also opendev doesn't exercise zuul-base-jobs, i added that repo because "providing example base jobs" was a lingering to do item for the v3 release, but it was never clear to me how to recommend consuming it (i was assuming folks would fork it sort of like opendev does) | 20:27 |
| *** pcaruana has quit IRC | 20:29 | |
| fungi | i wonder if there's a good way to test some standardized direct usage of that repo | 20:30 |
| SpamapS | Thanks folks | 20:32 |
| SpamapS | The team is unblocked and I can breathe again. I apologize (again) for the drama. | 20:32 |
| SpamapS | I do think the right thing to do is not use the git driver at all, and gate those repos. | 20:32 |
| SpamapS | (so, when a change lands in zuul-jobs or zuul-base-jobs, I should spin up a zuul and see if it can land a change. If I do it generically enough, we can include such a thing in zuul-jobs for everyone to be able to consume. | 20:33 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add add-fileserver test job https://review.opendev.org/670201 | 20:33 |
| fungi | SpamapS: that sounds awesome | 20:33 |
| fungi | and we probably have most of those pieces already | 20:34 |
| fungi | but important takeaway will be to recommend consuming those repos the same way the test does, i guess | 20:34 |
| fungi | because now there's probably at least several ways you could conceivably consume them | 20:35 |
| SpamapS | It's not bad the way it is now. Just that I seem to have chosen the set of options for running zuul that results in the most breakages, and I would like to know how we, as a project, want folks to use them. | 20:36 |
| fungi | i too would very much like to know that | 20:36 |
| SpamapS | I probably ignored several warning labels and caveats when doing it this way. | 20:36 |
| fungi | i regularly ignore six warning labels before breakfast | 20:37 |
| corvus | fungi: including "Do not eat" ? | 20:38 |
| fungi | technically that one is *during* breakfast | 20:40 |
| *** jamesmcarthur has joined #zuul | 20:42 | |
| mordred | fungi: is the warning to not eat an additional meal while eating breakfast? | 20:57 |
| fungi | since i ignored it, i don't really know | 21:06 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add add-gpgkey test job https://review.opendev.org/670206 | 21:13 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add add-launchpad-credentaials test job https://review.opendev.org/670207 | 21:14 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add add-sshkey test job https://review.opendev.org/670208 | 21:14 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add dstat graph roles https://review.opendev.org/670402 | 21:18 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Add mirror_info documentation https://review.opendev.org/669948 | 21:23 |
| *** tosky has quit IRC | 21:50 | |
| *** jamesmcarthur has quit IRC | 22:04 | |
| *** andreykurilin has quit IRC | 22:06 | |
| *** sgw has quit IRC | 23:01 | |
| *** sgw has joined #zuul | 23:18 | |
| *** jeliu_ has quit IRC | 23:30 | |
| *** mattw4 has quit IRC | 23:32 | |
| *** jamesmcarthur has joined #zuul | 23:35 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!