| *** harlowja has quit IRC | 00:23 | |
| *** dmellado has quit IRC | 01:22 | |
| *** gouthamr has quit IRC | 01:23 | |
| *** rlandy has quit IRC | 02:25 | |
| *** gouthamr has joined #zuul | 02:36 | |
| *** dmellado has joined #zuul | 02:38 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Use doc bindep for tox-docs https://review.openstack.org/594792 | 04:12 |
|---|---|---|
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul master: Update docs build job https://review.openstack.org/594510 | 04:12 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Update tox-docs to follow build-sphinx better https://review.openstack.org/594792 | 04:48 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Update tox-docs to follow build-sphinx better https://review.openstack.org/594792 | 04:57 |
| *** sdake has quit IRC | 05:13 | |
| *** sdake has joined #zuul | 05:13 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Update tox-docs to follow build-sphinx better https://review.openstack.org/594802 | 05:20 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul master: Update docs build job https://review.openstack.org/594510 | 05:27 |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Update tox-docs to follow build-sphinx better https://review.openstack.org/594802 | 05:34 |
| *** sdake has quit IRC | 05:50 | |
| *** sdake has joined #zuul | 05:51 | |
| openstackgerrit | Andreas Jaeger proposed openstack-infra/zuul-jobs master: Update tox-docs to follow build-sphinx better https://review.openstack.org/594802 | 05:55 |
| *** sdake has quit IRC | 06:02 | |
| *** sdake has joined #zuul | 06:04 | |
| *** pcaruana has joined #zuul | 06:41 | |
| *** dmellado has quit IRC | 07:04 | |
| *** dmellado has joined #zuul | 07:06 | |
| *** dmellado has quit IRC | 07:12 | |
| *** sshnaidm|afk is now known as sshnaidm | 07:17 | |
| *** dmellado has joined #zuul | 07:19 | |
| *** hashar has joined #zuul | 07:23 | |
| *** jpena|off is now known as jpena | 07:47 | |
| *** DangerousDaren has joined #zuul | 08:01 | |
| *** electrofelix has joined #zuul | 08:16 | |
| *** pcaruana has quit IRC | 08:28 | |
| *** pcaruana has joined #zuul | 08:30 | |
| *** jimi|ansible has quit IRC | 08:58 | |
| *** ttx has joined #zuul | 09:17 | |
| *** neilsun has joined #zuul | 09:18 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: web: enable build query by log_url https://review.openstack.org/594920 | 09:57 |
| *** jesusaur has quit IRC | 10:38 | |
| *** jesusaur has joined #zuul | 10:42 | |
| *** pbrobinson has left #zuul | 10:53 | |
| *** jpena is now known as jpena|lunch | 11:20 | |
| openstackgerrit | Merged openstack-infra/zuul-sphinx master: Update docs build job https://review.openstack.org/594518 | 11:32 |
| openstackgerrit | Merged openstack-infra/zuul-base-jobs master: Update docs build job https://review.openstack.org/594521 | 11:34 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Update tox-docs to follow build-sphinx better https://review.openstack.org/594802 | 11:37 |
| openstackgerrit | Merged openstack-infra/nodepool master: Update docs build job https://review.openstack.org/594517 | 11:41 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Update docs build job https://review.openstack.org/594522 | 11:48 |
| AJaeger | zuul team, I pushed https://review.openstack.org/594936 to redirect some docs.openstack.org/infra/zuul-* repos to zuul-ci.org, could you sanity check that, please? | 11:59 |
| *** ssbarnea|ruck has quit IRC | 12:13 | |
| *** jpena|lunch is now known as jpena | 12:20 | |
| *** rlandy has joined #zuul | 12:32 | |
| *** jimi|ansible has joined #zuul | 13:05 | |
| *** chmouel has joined #zuul | 13:28 | |
| openstackgerrit | Merged openstack-infra/zuul master: Update docs build job https://review.openstack.org/594510 | 13:40 |
| *** samccann has joined #zuul | 13:49 | |
| openstackgerrit | Filippo Inzaghi proposed openstack-infra/zuul-jobs master: TEST https://review.openstack.org/595011 | 13:55 |
| *** gouthamr has quit IRC | 14:00 | |
| *** dmellado has quit IRC | 14:01 | |
| *** chmouel has quit IRC | 14:09 | |
| *** jiapei has joined #zuul | 14:16 | |
| *** chmouel has joined #zuul | 14:16 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul-jobs master: Create tox_revoke_sudo variable for tox job. https://review.openstack.org/594569 | 14:42 |
| pabelanger | AJaeger: tristanC: jhesketh: corvus: mordred: updated ^ for new tox_revoke_sudp option, if you don't mind looking again. To compare, https://review.openstack.org/593150/ is another approach, which would result in using with-sudo in job name. | 14:43 |
| mnaser | pabelanger: i totally needed that yesterday lol | 14:43 |
| pabelanger | tox_revoke_sudo will result in less zuul jobs | 14:44 |
| pabelanger | mnaser: yah, please look then | 14:44 |
| mnaser | yeah i had to make my own thing to have the play without revoke-sudo | 14:44 |
| mnaser | i like the variable approach more | 14:45 |
| pabelanger | yah, I think tox_revoke_sudo will be the easier thing to use, but could mean jobs in PTI have the ability to use sudo now. | 14:45 |
| tristanC | pabelanger: can't the tox pti jobs force the revoke_sudo to true in the run playbook? | 14:48 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul master: DNM - testing docs https://review.openstack.org/595070 | 14:48 |
| pabelanger | tristanC: I don't think it can, without modifying the playbook. Then we need a new job | 14:48 |
| pabelanger | but haven't tested | 14:49 |
| *** gouthamr has joined #zuul | 14:50 | |
| corvus | the reason revoke-sudo is baked into there is that it's a really really bad idea to have a unit test job that uses sudo. people in openstack have done it before not realizing that means that other developers may be unwittingly destroying their workstations. the whole idea of using tox for unit tests is that it's easy for a developer to run before pushing up a change, but if it's not safe to do that, it | 14:53 |
| corvus | defeats the purpose. | 14:53 |
| corvus | it's fine to have a zuul job that uses sudo, but it shouldn't be easy to do that for a standard unit test job. | 14:53 |
| corvus | so openstack is going to specify in its PTI that projects use a job for unit tests which does not allow sudo, even if that means that openstack has to use a custom job for unit tests. | 14:54 |
| *** chmouel has quit IRC | 14:55 | |
| *** mmedvede is now known as mmedvede_ | 14:55 | |
| *** mmedvede_ is now known as mmedvede | 14:55 | |
| corvus | but to the extent that zuul-jobs is trying to incorporate best practice, we may want to consider keeping the standard tox-pyXX jobs forcing revoke sudo, and adding alternate jobs which allow it. | 14:55 |
| pabelanger | in that case, that's what tox-with-sudo does: https://review.openstack.org/593150/ leaves the existing tox job alone and starts a new one for other jobs to parent too | 15:00 |
| *** chmouel has joined #zuul | 15:01 | |
| pabelanger | and if we need py36 with sudo, per tristanC comment. We could add tox-py36-with-sudo job into zuul-jobs. | 15:01 |
| corvus | pabelanger: we could, but that sounds like a bad idea. i think the use of sudo should be constrained to special jobs, not regular unit tests. | 15:02 |
| mnaser | corvus: in that case, how do you feel about things like doing functional testing of something like an ansible role | 15:03 |
| corvus | mnaser: i think it's great! | 15:03 |
| mnaser | but we're using something like tox -efunctional to do that, so then is there an alternate way (i'm open) | 15:03 |
| corvus | i want to be really clear here, so i'll repeat myself -- using sudo is FINE! lots of jobs should do it! just not unit test jobs. | 15:03 |
| mnaser | oh yes i agree | 15:03 |
| mnaser | we run linters without sudo, but functionals with.. but i dunno, it doesnt feel so nice :< | 15:04 |
| corvus | mnaser: 'tox -e functional' sounds great to me, is there a downside? | 15:04 |
| mnaser | corvus: one job needs sudo and another doesn't, though i dont think functional tests of ansible roles are possible without sudo unfortunately | 15:04 |
| corvus | i'm still not seeing the problem :( | 15:05 |
| * rcarrillocruz TIL about PTI pages... | 15:07 | |
| *** pcaruana has quit IRC | 15:10 | |
| pabelanger | tox -efunctional is also a fine job name | 15:14 |
| *** openstackgerrit has quit IRC | 15:31 | |
| *** chmouel has quit IRC | 15:32 | |
| *** dmellado has joined #zuul | 15:45 | |
| SpamapS | {name: tox-efunctional, parent: tox-with-sudo, vars: { environment: functional }} | 15:49 |
| *** pcaruana has joined #zuul | 15:49 | |
| SpamapS | i done wrote a job | 15:50 |
| *** gouthamr has quit IRC | 16:05 | |
| *** jpena is now known as jpena|off | 16:09 | |
| SpamapS | tristanC: FYI, the latest aws nodepool driver patch is 'sploding for me | 16:21 |
| SpamapS | http://paste.openstack.org/show/728615/ | 16:21 |
| SpamapS | Kinda looks like maybe the tests aren't actually exercising aws | 16:22 |
| * SpamapS is looking into it now | 16:22 | |
| SpamapS | (pretty sur ethe fix is easy tho) | 16:23 |
| AJaeger | SpamapS: http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/zuul.d/jobs.yaml#n430 | 16:24 |
| *** hashar has quit IRC | 16:34 | |
| *** jiapei has quit IRC | 16:35 | |
| *** gouthamr has joined #zuul | 16:46 | |
| *** neilsun has quit IRC | 17:05 | |
| *** hashar has joined #zuul | 17:09 | |
| *** elyezer has quit IRC | 17:10 | |
| *** elyezer has joined #zuul | 17:11 | |
| *** electrofelix has quit IRC | 17:25 | |
| *** mmedvede has quit IRC | 17:25 | |
| *** jpena|off has quit IRC | 17:25 | |
| *** panda|off has quit IRC | 17:25 | |
| *** EmilienM has quit IRC | 17:25 | |
| *** jlvillal has quit IRC | 17:25 | |
| *** Diabelko has quit IRC | 17:25 | |
| *** corvus has quit IRC | 17:25 | |
| *** jlvillal has joined #zuul | 17:27 | |
| *** hashar has quit IRC | 17:27 | |
| *** gouthamr has quit IRC | 17:29 | |
| *** jeblair has joined #zuul | 17:30 | |
| *** EmilienM has joined #zuul | 17:30 | |
| *** jpena has joined #zuul | 17:36 | |
| *** jeblair is now known as corvus | 17:41 | |
| *** elyezer has quit IRC | 17:43 | |
| SpamapS | AJaeger: ^^ mnaser | 17:47 |
| SpamapS | AJaeger: rather, thanks.. ^^ was for mnaser | 17:47 |
| SpamapS | mmm just made 3 new executors live. | 17:48 |
| * SpamapS pets zuul... good service. | 17:48 | |
| *** DangerousDaren has quit IRC | 17:48 | |
| Shrews | SpamapS: careful... it bites | 17:49 |
| jlk | I might actually get to do some Zuul hacking this weekend, while on a houseboat with my dad and family. | 17:51 |
| jlk | Working on github3.py app auth support and checks API | 17:51 |
| SpamapS | neat | 17:57 |
| *** mmedvede has joined #zuul | 18:00 | |
| corvus | jlk: that's great! i added line comment support in the gerrit driver, hopefully we can do the same with the checks api | 18:03 |
| jlk | corvus: that's the plan! | 18:03 |
| jlk | https://github.com/sigmavirus24/github3.py/pull/888 is the open PR for adding the library interface to support these APIs | 18:03 |
| jlk | which will make doing it in Zuul quite easier. | 18:04 |
| corvus | slick | 18:04 |
| mnaser | SpamapS: heh, nice | 18:06 |
| mnaser | i was wondering if we should drop our local zuul-jobs and consume 'zuul-jobs' directly | 18:06 |
| *** ChanServ has quit IRC | 18:16 | |
| mnaser | (is that something that's okay in general ^) | 18:22 |
| *** ChanServ has joined #zuul | 18:22 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 18:22 | |
| *** dmellado has quit IRC | 18:22 | |
| *** elyezer has joined #zuul | 18:34 | |
| AJaeger | corvus: should we link on zuul-ci to zuul-sphinx? Please see https://review.openstack.org/591282 | 18:58 |
| corvus | AJaeger: lgtm. we're wildly inconsistent with our trailing slashes! | 19:08 |
| corvus | of course, now we need a 6th piece of documentation to balance out the page again | 19:09 |
| AJaeger | ;) | 19:09 |
| dmsimard | There's no mention of "User authorization callback URL" in the docs for GitHub setup: https://zuul-ci.org/docs/zuul/admin/github_setup.html | 19:09 |
| dmsimard | But it's a required field when creating a GitHub app.. does anyone know what it should be ? | 19:10 |
| AJaeger | corvus: https://review.openstack.org/#/c/592292/ merged, so now docs.openstack.org/infra/zuul-sphinx redirects to zuul-ci.org/docs/zuul-sphinx | 19:10 |
| AJaeger | oops, did not publish yet | 19:10 |
| AJaeger | yeah, too large backlog... | 19:10 |
| *** openstackgerrit has joined #zuul | 19:10 | |
| openstackgerrit | Merged openstack-infra/zuul-website master: Link to Zuul Sphinx extension https://review.openstack.org/591282 | 19:10 |
| dmsimard | https://i.imgur.com/IOM6gcn.png | 19:11 |
| corvus | AJaeger: https://zuul-ci.org/docs/zuul-sphinx/ exists | 19:12 |
| corvus | AJaeger: oh, you mean the redirect isn't in place yet. gotcha | 19:12 |
| AJaeger | corvus: exactly, 4h backlog of post queue. Let's check tomorrow for the redirect. Content is available in both right now and we don't want anybody to find outdated docs on docs.o.o/infra/... | 19:12 |
| jlk | dmsimard: hrm. One sec | 19:17 |
| SpamapS | hm | 19:18 |
| SpamapS | anyone know why I might be seeing this: http://paste.openstack.org/show/728629/ | 19:18 |
| SpamapS | set up some new executors and they don't seem to be able to merge | 19:19 |
| SpamapS | ValueError: SHA b'2d5c655650e86201e0636f8a9105d2992c36340c' could not be resolved, git returned: b'2d5c655650e86201e0636f8a9105d2992c36340c missing' | 19:19 |
| jlk | "In "User authorization callback URL", type the full URL to redirect to after a user authorizes an installation. This URL is used if your app needs to identify and authorize user-to-server requests." | 19:19 |
| jlk | dmsimard: I don't believe Zuul does any user-to-server things, so this could just point to the Zuul install's home page | 19:20 |
| dmsimard | jlk: should we add that to the docs to remove the ambiguity ? | 19:20 |
| jlk | Yeah if it's a confusing thing, which it seems to be | 19:20 |
| jlk | https://developer.github.com/apps/building-github-apps/identifying-and-authorizing-users-for-github-apps/ | 19:21 |
| jlk | that's what it's talking about | 19:21 |
| SpamapS | seems like the initial jobs that ran on this executor cloned the repos, but now they're not fetching | 19:22 |
| corvus | SpamapS: it looks like there may be some ref that's on a merger (or executor) that isn't on that executor. what's the line before that one? | 19:24 |
| corvus | SpamapS: (in the debug log if you have it) | 19:24 |
| SpamapS | http://paste.openstack.org/show/728630/ | 19:25 |
| SpamapS | there's a bunch more lines before | 19:25 |
| SpamapS | now, I'd expect that at some point it needs to go fetch that ref from github | 19:26 |
| SpamapS | (I do not have any dedicated mergers) | 19:26 |
| corvus | SpamapS: the 'Updating repository' line should have done that, it calls this: https://git.openstack.org/cgit/openstack-infra/zuul/tree/zuul/merger/merger.py#n368 | 19:29 |
| corvus | SpamapS: well, that should have fetched the tags from the executor's cache. previously, the executor should have updated its cache from its upstream (presumably github) in the same way. | 19:30 |
| corvus | SpamapS: but this line, which precedes the error, is the ref / commit that the error is about: 2018-08-22 12:24:32,484 DEBUG zuul.AnsibleJob: [build: ea977a3d26a3447bb3bcc79b9a510be4] Create reference refs/tags/deploy0 at 2d5c655650e86201e0636f8a9105d2992c36340c in /tmp/tmplohxe_ka/ea977a3d26a3447bb3bcc79b9a510be4/work/src/github.secureserver.net/cloudplatform/hoist | 19:31 |
| corvus | SpamapS: is there, by any chance, a possibily that refs/tags/deploy0 was deleted from upstream? | 19:31 |
| SpamapS | corvus: not that I know of | 19:32 |
| SpamapS | actually I don't see that tag | 19:33 |
| SpamapS | not even sure where it's getting that tag from | 19:33 |
| corvus | SpamapS: if it existed on the merger node (in your case, some random executor) at the time the initial merge was performed (shortly after the item was enqueued) and then removed between then and the time the job started, i would expect to see that error | 19:34 |
| SpamapS | I can't find that tag anywhere | 19:35 |
| SpamapS | or even that string | 19:35 |
| SpamapS | in any configs or anything | 19:35 |
| SpamapS | but I think I see what may be breaking stuff | 19:36 |
| SpamapS | that build was started on an executor which I stopped | 19:36 |
| * SpamapS tries restarting the old one | 19:36 | |
| corvus | i have to grab lunch; biab | 19:37 |
| SpamapS | corvus: ty for looking | 19:38 |
| *** pcaruana has quit IRC | 19:47 | |
| *** weshay has joined #zuul | 19:47 | |
| weshay | any thoughts on how to checkout out an upstream change and automatically pull any chain of deps from a depends-on? | 19:48 |
| weshay | locally outside of the upstream ci system | 19:48 |
| *** rfolco has joined #zuul | 19:48 | |
| weshay | mordred, ^ | 19:51 |
| *** panda has joined #zuul | 19:55 | |
| *** harlowja has joined #zuul | 20:42 | |
| SpamapS | weshay: that wouldn't be a super hard entry-point to write into zuul | 20:43 |
| SpamapS | like zuul-checkout <connectionname> <project> <change-#> <job-name> | 20:43 |
| weshay | rlandy, ^ | 20:43 |
| SpamapS | and have it assemble the required-projects from that job with that change the same way the executor does | 20:43 |
| SpamapS | the only sort of wonky thing would be configuring the connections on your local machine | 20:44 |
| SpamapS | but yeah, I think you could write that as an addition to zuul fairly easily | 20:44 |
| SpamapS | (to be clear, zuul-checkout doesn't exist ;) | 20:44 |
| SpamapS | it used to | 20:44 |
| SpamapS | but things are different now | 20:45 |
| rlandy | off hand, do you know where we would look for the code where the executor figures this (zuul_changes) out? to comapre | 20:46 |
| weshay | SpamapS, ya.. I thought that's what zuul-cloner was supposed to do, but that project died | 20:46 |
| SpamapS | zuul-cloner relies on ZUUL_REF | 20:49 |
| SpamapS | which isn't a thing when the mergers don't expose refs anymore. | 20:49 |
| SpamapS | I was just looking .. I think it's going to be mostly in zuul.merger | 20:51 |
| weshay | thanks SpamapS | 20:56 |
| *** jpena is now known as jpena|off | 20:57 | |
| SpamapS | yeah looks like if you can just instantiate a Repo object that gets you pretty far, and then you might be able to refactor some of the stuff in zuul.manager to be consumable for assembling the dependencies | 21:01 |
| SpamapS | (I've said a few times that I'd really love to have a localized CLI zuul-executor so keep me posted if you get some traction) | 21:02 |
| *** samccann has quit IRC | 21:07 | |
| corvus | there are some stories in storyboard which describe the problem and solution | 21:08 |
| *** gouthamr has joined #zuul | 21:12 | |
| openstackgerrit | Jarek Lukow proposed openstack-infra/zuul master: Add missing parentheses to title invocation https://review.openstack.org/595361 | 21:21 |
| mnaser | i asked this earlier but might have been missed | 21:25 |
| mnaser | should we use upstream openstack-infra/zuul-jobs or maintain a local identical clone | 21:25 |
| corvus | mnaser: we aim to support zuul-jobs continuously; the policies around new features and deprecation are designed to support that. it should be fine if you upgrade zuul relatively soon after each release (or, obviously, CD it). but if you don't want to commit to that or just want the extra safetly of pulling on your schedule, of course that's fine | 21:34 |
| mnaser | corvus: i see, well, we're involved enough to get a grasp on whats going on but i guess the important thing here is it's not a movable target that can get super fundamental changes overnight | 21:35 |
| corvus | mnaser: right. super-fundamental changes take at least a few weeks. :) https://zuul-ci.org/docs/zuul-jobs/deprecation.html | 21:37 |
| corvus | does anyone happen to have a role that writes out a near-copy of the zuul inventory as a plain ansible inventory? specifically, i just need the hosts and ip addresses, but not the zuul user, etc. | 21:37 |
| corvus | odyssey4me, SpamapS: ^ perhaps? | 21:37 |
| SpamapS | corvus: we still have it written as a yaml file | 21:38 |
| corvus | SpamapS: oh yeah that's fine | 21:38 |
| corvus | SpamapS: i basically just need to strip out all the extra stuff | 21:38 |
| SpamapS | Whatever mine does, it happens via zuul-jobs roles. | 21:39 |
| corvus | SpamapS: hrm, i didn't see a zuul-jobs role to do this | 21:39 |
| * SpamapS digs through | 21:44 | |
| dmsimard | SpamapS, tristanC: btw I added comments on the nodepool AWS driver patch https://review.openstack.org/#/c/535558/ I'm working on a poc and I have a test environment readily available for... testing | 22:30 |
| *** ChanServ has quit IRC | 22:49 | |
| SpamapS | corvus: sorry I'm having trouble finding where we actually write it out | 22:50 |
| corvus | SpamapS: ok, thanks. i'm writing a nifty new role | 22:51 |
| SpamapS | corvus: excellent. :) | 22:54 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add write-inventory role https://review.openstack.org/595378 | 23:03 |
| *** ChanServ has joined #zuul | 23:03 | |
| *** barjavel.freenode.net sets mode: +o ChanServ | 23:03 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: Add write-inventory role https://review.openstack.org/595378 | 23:32 |
| *** rlandy has quit IRC | 23:39 | |
| *** gouthamr has quit IRC | 23:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
