| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Rename zuul-revoke-sudo to revoke-sudo https://review.openstack.org/481331 | 00:15 |
|---|---|---|
| mordred | jeblair: ok - that topic just got a bit longer | 00:15 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Rename git_host to server in github driver https://review.openstack.org/477594 | 00:59 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Clarify canonical_hostname documentation https://review.openstack.org/479020 | 00:59 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add docs on allow-secrets https://review.openstack.org/480726 | 00:59 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix indentation error in docs https://review.openstack.org/480740 | 00:59 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Move status_url to webapp config section https://review.openstack.org/480759 | 01:00 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Pass result data back from ansible https://review.openstack.org/479442 | 01:05 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: WIP: remove success/failure url options https://review.openstack.org/481342 | 01:05 |
| openstackgerrit | Jamie Lennox proposed openstack-infra/zuul feature/zuulv3: Cleanup on ssh-agent failure https://review.openstack.org/481344 | 01:06 |
| jeblair | mordred: 481327 has an error | 01:07 |
| *** jkilpatr has quit IRC | 01:08 | |
| jamielennox | signed ssh keys are cool, however the infrastructure around them is a lot of effort. not just revocations but nodepool would need to reach out somehow to ask for keys to be signed and you would need to maintain that list as a service | 01:09 |
| jamielennox | deploying something like vault would give you a similar effect with per host ssh keys and the auditting built in, it's just whether it's worth the effort | 01:14 |
| jamielennox | having the private key there in zookeeper as a one use is fairly good for this, it's just annoying that you have to actually upload this key to nova before booting the vm | 01:15 |
| clarkb | jamielennox: well I think we are saying that system is also too clunky | 01:31 |
| clarkb | jamielennox: and maybe having zuul swap out keys is best | 01:32 |
| clarkb | and ya ldap can have ssh keys too its a matter of how much extra infrastructure and/or tooling is reasonable | 01:33 |
| clarkb | seems like zuul can bootstrap this itself with a minimal level of effort | 01:33 |
| jamielennox | yea, so that's the process now right, there's the built in key that zuul logs in with and swaps in a per job key, which is pretty good for now | 01:33 |
| mordred | ++ | 01:34 |
| jamielennox | per build would get that better, but ugh, don't want to have to maintain ldap just for zuul if i can avoid it | 01:36 |
| fungi | jeblair: having read through both stories, i think i have a marginal preference for the required/immutable base job using the primary shared key solely for bootstrapping a secondary per-build key. this gets nodepool out of the picture and also avoids any scalability concerns around nova keypair quotas | 02:05 |
| fungi | catching up on the subsequent conversation, it looks like maybe others came to similar conclusions | 02:07 |
| fungi | i also like that having zuul bootstrap the secondary key reduces the extent to which key material is being shipped around between different parts of the system. every move of that data is another risk, in the end | 02:10 |
| tristanC | fungi: mordred: the sql schema likely needs to be upgrade, e.g. it's not well designed for periodic/post pipeline that doesn't have change number/newrev | 02:26 |
| tristanC | it's also missing a tenant column | 02:27 |
| tristanC | then it seems relatively easy to build a openstack-health like service, either by direct database query, or through a zuul-web controller | 02:29 |
| tobiash | jeblair: I'm also in favor of the zuul base job story | 05:14 |
| tobiash | jeblair: we just should make sure that the base job also handles nodeless jobs gracefully | 05:14 |
| tobiash | jeblair: I'd like to retain the ability to run lightweight local-only jobs on the executor (e.g. for mirroring jobs) | 05:15 |
| tobiash | jeblair: I especially like the automatic base job inheritance as this would prevent the user from adding a job without inheriting and then wondering why he doesn't get logs :) | 05:33 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Optionally bind /lib64 https://review.openstack.org/481432 | 05:35 |
| *** bhavik1 has joined #zuul | 05:44 | |
| *** ianw is now known as ianw_pto | 05:55 | |
| *** ajafo has quit IRC | 06:26 | |
| *** ajafo has joined #zuul | 06:26 | |
| *** pabelanger has quit IRC | 06:37 | |
| *** yolanda_ has quit IRC | 06:38 | |
| *** cinerama has quit IRC | 06:38 | |
| *** pabelanger has joined #zuul | 06:40 | |
| *** cinerama has joined #zuul | 06:40 | |
| *** yolanda_ has joined #zuul | 06:44 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Add status_pattern to zuul-web config https://review.openstack.org/481474 | 07:33 |
| *** bhavik1 has quit IRC | 08:02 | |
| *** hashar has joined #zuul | 08:59 | |
| *** openstackgerrit has quit IRC | 09:48 | |
| *** hashar is now known as hasharLunch | 10:37 | |
| mordred | tobiash: ++ | 11:22 |
| *** openstackgerrit has joined #zuul | 11:29 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Remove the old tox-linter job https://review.openstack.org/481322 | 11:29 |
| tobiash | tried out Shrews websocket based log streaming patch now in my v3 deployment and it works :) | 11:30 |
| mordred | tobiash: woot! | 11:30 |
| mordred | tobiash: did you try it with a command line websocket client or with an html one? | 11:31 |
| tobiash | mordred: I used a html one | 11:31 |
| mordred | \o/ | 11:34 |
| mordred | jeblair: I think https://review.openstack.org/#/c/481586/ may be exposing a bug in job config :( | 11:36 |
| mordred | jeblair: that is trying to create a job in openstack-zuul-jobs that has a job in zuul-jobs as its parent | 11:36 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove old tox jobs https://review.openstack.org/481325 | 11:38 |
| mordred | jeblair: I have verified that zuul-tox-docs is in zuul-jobs - so I'm concerned that parent: zuul-tox-docs is not working in openstack-zuul-jobs | 11:39 |
| mordred | tristanC: ++ to upgrading sql schema | 11:48 |
| mordred | tristanC: oh my. there are no indexes in the schema either | 11:49 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Improve debugging at start of job https://review.openstack.org/480761 | 11:51 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: github: gracefully handle unknown event https://review.openstack.org/473249 | 11:53 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Add html based websocket client for console stream https://review.openstack.org/481599 | 11:54 |
| tobiash | mordred: this is the html based websocket client I used ^^ | 11:55 |
| mordred | tobiash: fantastic! thank you | 12:01 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add some indexes to allow efficient dashboard sql https://review.openstack.org/481614 | 12:37 |
| mordred | tristanC: ^^ that should allow for some better dashboard queries :) | 12:38 |
| *** hasharLunch is now known as hashar | 12:39 | |
| mordred | tristanC: also - we should think a bit about github changes + gerrit changes - change is a unique id for gerrit - but not for github - which means project+change is the unique identifier | 12:43 |
| tristanC | mordred: maybe the buildset table also needs a source trigger field | 12:46 |
| mordred | tristanC: so if we wanted a view in the dashboard which is "show me all the jobs for this change" - it would wind up needing to be something like /dashboard/#!/change/review.openstack.org/openstack-infra/zuul/481614 or /dashboard/#!/change/github.com/j2sol/z8s-sandbox/2 | 12:46 |
| mordred | tristanC: that is - if we wanted such a view to be guessable by the user | 12:46 |
| mordred | tristanC: if we reorged that slightly --- and did /dashboard/#!/review.openstack.org/openstack-infra/zuul/change/481614 - then /dashboard/#!/review.openstack.org/openstack-infra/zuul could be a page to show the jobs for a given project - and both could be urls a user could be expected to guess | 12:48 |
| mordred | tristanC: and then /dashboard/#!/jobs and /dashboard/#!/jobs/tox-py27 would make sense | 12:50 |
| mordred | tristanC: but jobs view by project maybe /dashboard/#!/review.openstack.org/openstack-infra/zuul/jobs ? | 12:51 |
| tristanC | mordred: heh I meant to ask, can a job be defined differently in two projects? | 12:52 |
| mordred | tristanC: a project can define different variants | 12:52 |
| mordred | tristanC: so a project can say "please run my tox-py27 job for the stable/mitaka branch on centos-7 instead of ubuntu-xenial" | 12:53 |
| mordred | tristanC: but it'll still be "the tox-py27 job" | 12:53 |
| mordred | tristanC: heh. we probably want to put branch in here too | 12:54 |
| mordred | tristanC: there's definitely a difference in "I want to see all jobs" "I want to see jobs for stable/ocata" :) | 12:54 |
| tristanC | wouldn't using optional query-string be easier to manage then? e.g. /jobs/?project=$name&branch=$name ... ? | 12:55 |
| tristanC | and similarly, /builds/?project=$name&pipeline=$name | 12:56 |
| mordred | tristanC: maybe? I was mostly basing those urls on my (limited) understanding of how people like to do url construction in angular-land | 12:56 |
| tristanC | mordred: heh, I never build a complex enough angular thing to need such construction :-) | 12:57 |
| mordred | :) | 12:57 |
| mordred | gerrit does a similar: https://review.openstack.org/#/q/status:open+project:openstack-infra/zuul+branch:feature/zuulv3 | 12:58 |
| mordred | but that's for arbitrary searching | 12:58 |
| tristanC | though I'm more thinking about the rest controller and how the route would map in aiohttp | 12:58 |
| mordred | ++ | 12:58 |
| mordred | yah- I think that's the more important thing to consider | 12:58 |
| mordred | tristanC: and in that perspective, yah - a /builds and a /jobs endpoint seem much more sane | 13:01 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Optionally bind /lib64 https://review.openstack.org/481432 | 13:34 |
| tobiash | mordred: what do you think about adding empty lines before the plays in the build logs? | 13:57 |
| tobiash | mordred: I think that would make them more readable (maybe also for the tasks) | 13:58 |
| mordred | tobiash: I think it's a good idea - I actually want to run an html post-processor on the build logs before we upload them ... | 14:00 |
| mordred | tobiash: that would transform them into something like this: http://logs.openstack.org/82/478582/2/check/gate-shade-python35/9d193c8/console.html | 14:00 |
| mordred | (in fact, i just wnat to use the code that does that transform) | 14:00 |
| mordred | tobiash: so I think adding a blank line before a play header wouldn't be a terrible choice - or even just adding a css selector onto lines with play in them so that a local css could be provided that formats with extra padding, etc | 14:01 |
| tobiash | mordred: sounds also good | 14:01 |
| mordred | tobiash: I like your optionally-bind patch - I kind of think we should go one step further, have there be a list of dirs to ro-bind and have all of them be optional bindings only if they exist | 14:03 |
| mordred | (well, I think ssh_auth_sock needs to be hard-coded and not in the list of ro-binds) | 14:03 |
| mordred | I'm betting we'll find a few more platform-specific dirs we want to bind or not bind | 14:04 |
| tobiash | mordred: I also thought about that but thought the remaining ro dirs are pretty standard and don't need checking | 14:06 |
| tobiash | mordred: but I can change that as well (just wanted to do a minimal impact patch) | 14:07 |
| jeblair | mordred: i know of no reason why the zuul-tox-docs thing should not work, so i will investigate that as a bug immediately. | 14:10 |
| tobiash | mordred, jeblair: because of missing repo update? | 14:13 |
| tobiash | mordred, jeblair: noticed that the playbook repo update patch did not solve the issue of not updating to latest origin/master | 14:14 |
| tobiash | but did not yet have time to investigate | 14:14 |
| jeblair | tobiash: oh, hrm. i will look into that. | 14:14 |
| tobiash | mordred, jeblair: just guessing, but the sentence 'job should fail/notfail but doesn't' made me think that might be the same issue | 14:17 |
| jeblair | did we have a story about a problem with new patchsets? | 14:21 |
| tobiash | jeblair: not that I know of | 14:26 |
| jeblair | http://paste.openstack.org/show/614780/ | 14:28 |
| jeblair | i'm confused by that error which i'm seeing in our logs | 14:28 |
| jeblair | _executeJobs is called from executeJobs which starts with the line: "if not item.current_build_set.layout: return False" | 14:29 |
| jeblair | so i have no idea how item.current_build_set.layout could be None in that traceback | 14:30 |
| jeblair | as i'm looking into this, i notice that we perform a merge operation for every non-active item in the check pipeline. i don't think we need to do that, i think we can do them only for active items. i'll note that for a future optimization. | 15:00 |
| jeblair | s/active/live/ | 15:01 |
| *** hashar is now known as hasharAway | 15:45 | |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul feature/zuulv3: Ensure github change body is not a NoneType https://review.openstack.org/481681 | 15:50 |
| jlk | jeblair: SpamapS: mordred: I have a small stack of github fixes, starting with https://review.openstack.org/481314 that could use some reviews/merges. | 15:51 |
| jlk | Does it seem a little ironic that just as Zuul is walking away from it's use of jenkins job builder, there is a big push to create a JJB community and future development? What else in openstack land uses jjb? | 16:06 |
| jeblair | jlk: jjb is widely used outside of openstack, so i think it's great. :) | 16:08 |
| jlk | I think it's great too, but it's just odd that there won't be any consumers within the org that is hosting the project. | 16:09 |
| clarkb | we host a lot of software that we don't consume | 16:11 |
| jeblair | jlk: we're not going to pull the rug out from under it. we're happy to have them here as long as they want to stay. :) | 16:11 |
| mordred | yah. we also wouldn't stop them if they wanted to move elsewhere - but happy to keep them | 16:15 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Fix AttributeError in executor client https://review.openstack.org/481691 | 16:26 |
| jeblair | mordred: i miss compilers ^ :( | 16:26 |
| jeblair | that should be the fix for the error that currently has zuul stuck. now i'm going to try to figure out if that, or something else, is responsible for the error you reported. | 16:27 |
| mordred | jeblair: mine is very reproducible - we could land that fix and restart and trigger my patch again | 16:46 |
| mordred | clarkb, SpamapS: wanna do a quick +3: https://review.openstack.org/#/c/481691 ? | 16:46 |
| jeblair | mordred: well, the fact that your changes are stuck because of the 'item' error could be affecting that behavior, and restarting could paper over tobiash's theory about not correctly updating the config, so i want to poke a bit more before doing that | 16:48 |
| mordred | jeblair: ah - ok | 16:48 |
| jeblair | mordred: okay, i don't think either of those things should be affecting your change; so i'm going to +3 my bugfix and restart with it and see if we still see your error (i expect us to). | 16:55 |
| jeblair | mordred: according to the logs, it parsed exactly the config i would expect | 16:55 |
| mordred | jeblair: *awesome* | 16:58 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix AttributeError in executor client https://review.openstack.org/481691 | 17:03 |
| jeblair | mordred: okay, i restarted. i'm going to recheck some of your changes now | 17:16 |
| jeblair | yay, it reproduced | 17:17 |
| mordred | jeblair: yay! | 17:19 |
| mordred | jeblair: I like it when I can provide you with weird bugs | 17:19 |
| jeblair | i wrote a test case to mimic this and it passes | 17:19 |
| jeblair | mordred: aha! it's the .zuul.yaml -> zuul.yaml rename | 17:28 |
| jeblair | mordred: apparently there's a place where we don't read zuul.yaml from untrusted projects | 17:29 |
| mordred | jeblair: oh! well then | 17:30 |
| jeblair | it looks like we read either from config projects, but only .zuul.yaml from untrusted projects. if anything, that feels backwards. but maybe we should just read either from both kinds of projects. :) | 17:32 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Read config from .zuul.yaml or zuul.yaml https://review.openstack.org/481706 | 17:36 |
| jeblair | mordred: ^ | 17:36 |
| mordred | jeblair: yah- I liked your narrative description in the docs | 17:39 |
| mordred | "they're both valid - you may want to use the hidden file for repos where it's not the primary purpose" | 17:40 |
| mordred | jeblair: while your head has all of that context loaded in it - https://review.openstack.org/#/c/473764/ seems like it's in good shape | 17:41 |
| mordred | jeblair: heh. although amusingly enough it's going to merge-conflict on your patch - and it copied the logic in that place - so the merge conflict will _actually_ show a place where the logic got fixed in HEAD and should be similarly applied to the patch | 17:42 |
| jeblair | mordred: do you have any idea what changed between PS3 and current on that? | 17:43 |
| jeblair | i've diffed 3-4 4-5 and 5-6, and there are big rebases in all of them | 17:43 |
| mordred | jeblair: I do not | 17:44 |
| jeblair | so if i'm going to review that, i'm going to need to manually interdiff | 17:44 |
| jeblair | i'd love to land it, but i don't have time for that now | 17:44 |
| jeblair | i will leave comments asking for some review help | 17:44 |
| mordred | ++ | 17:45 |
| mordred | jlk: gh stack lgtm - maybe SpamapS can knock those in | 17:53 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Read config from .zuul.yaml or zuul.yaml https://review.openstack.org/481706 | 17:56 |
| mordred | jeblair: I'd like to restart zuul-executor with that ^^ applied - that gonna mess you up? | 18:11 |
| jeblair | mordred: nope; i'll go ahead and restart now | 18:36 |
| jeblair | mordred: done | 18:37 |
| jeblair | rechecked 327 | 18:38 |
| jeblair | mordred: +1! | 18:42 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Duplicate the zuul-tox jobs to just be named tox https://review.openstack.org/481323 | 19:09 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Consume tox instead of zuul-tox https://review.openstack.org/481326 | 19:50 |
| SpamapS | mordred: sorry, stomach bugged knocked me out of commission. :P | 19:50 |
| SpamapS | Just peeking at my IRC before I crawl back into hole. | 19:51 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Handle GitHub comment reviews carefully https://review.openstack.org/481314 | 19:58 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Remove unnecessary loop in github test https://review.openstack.org/481315 | 19:59 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Ensure github change body is not a NoneType https://review.openstack.org/481681 | 20:00 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Pass result data back from ansible https://review.openstack.org/479442 | 20:02 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Pass result data back from ansible https://review.openstack.org/479442 | 20:03 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Pass result data back from ansible https://review.openstack.org/479442 | 20:07 |
| mordred | jeblair: yay! | 20:42 |
| mordred | SpamapS: :( | 20:42 |
| mordred | jeblair: https://review.openstack.org/#/q/topic:rename-dance has three more patches you might reasonably have missed because WOW it's a lot of patches | 20:45 |
| jeblair | mordred: thanks, i hadn't missed them, but they had slipped my mind after rechecking them a little bit ago. will +3 now | 20:46 |
| mordred | jeblair: :) | 20:46 |
| mordred | jeblair: I just - I figure zuul is the only one who can ACTUALLY understand the depends-on graph | 20:46 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Remove the transitionary zuul-tox jobs https://review.openstack.org/481324 | 20:47 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Rename zuul-revoke-sudo to revoke-sudo https://review.openstack.org/481331 | 20:50 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Pass result data back from ansible https://review.openstack.org/479442 | 20:52 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Support relative urls in success-url https://review.openstack.org/481342 | 20:52 |
| jeblair | mordred: heh, probably :) | 20:52 |
| jeblair | mordred: those 2 changes are ready, i think. that's my attempt to make the log url stuff easy, but also keep the neat trick we do where we deep link to docs builds. | 20:53 |
| jeblair | that's far from the only approach we could take there... let me know what you think | 20:54 |
| mordred | jeblair: reading now | 20:54 |
| mordred | jeblair: what's the ZUUL_JOBDIR env var in the ansible module for? | 20:55 |
| mordred | jeblair: oh - that works? | 20:55 |
| mordred | I guess since it only ever gets run on localhost ... | 20:55 |
| jeblair | mordred: so it knows the right place to put results.json by default. you can set a 'path:' parameter, but i don't want folks to have to set it every time; it's only as an override. | 20:56 |
| mordred | yah | 20:56 |
| jeblair | mordred: exactly, yes. localhost only. | 20:56 |
| mordred | jeblair: so - fwiw, since this is intended to only ever run on localhost - you COULD implement this as an action plugin | 20:56 |
| jeblair | mordred: i tried using cwd, but it seems that it ends up being the playbook directory. | 20:56 |
| jeblair | mordred: oh interesting. that would give us access to facts, i assume. | 20:57 |
| mordred | jeblair: I'm not saying that would be better - but just more of from an fyi point of view | 20:57 |
| jeblair | then we could drop the env var. | 20:57 |
| jeblair | honestly, if so, i think that would be preferable. though maybe as a followup patch. | 20:58 |
| mordred | jeblair: ++ | 21:01 |
| *** jesusaur has quit IRC | 21:28 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add work_dir option to tox role https://review.openstack.org/481762 | 21:29 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add a job that runs zuul's py35 unit tests https://review.openstack.org/481763 | 21:29 |
| mordred | jeblair: ^^ so that should be fun | 21:29 |
| *** jesusaur has joined #zuul | 21:32 | |
| jeblair | mordred: that is really exciting. :) | 21:33 |
| mordred | jeblair: the error is even more exciting | 21:33 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add a job that runs zuul's py35 unit tests https://review.openstack.org/481763 | 21:35 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add work_dir option to tox and bindep roles https://review.openstack.org/481762 | 21:35 |
| mordred | jeblair: I made a param for the tox role - but not the bindep role - so bindep ran for zuul-jobs and then tox ran for zuul :) | 21:35 |
| jeblair | mordred: how did you deduce that from that error? :) | 21:37 |
| mordred | jeblair: 2017-07-07 21:30:00.047444 in http://logs.openstack.org/63/481763/1/check/f731d71/job-output.txt | 21:38 |
| jeblair | mordred: if pabelanger were here, he might want us to think twice about the var name -- is "work_dir" sufficiently scoped? should it be zuul_work_dir ? | 21:39 |
| jeblair | mordred: also | 21:41 |
| jeblair | 2017-07-07 21:36:32.041662 | TASK [test-setup : Run tools/test-setup.sh chdir=src/{{ zuul.project.canonical_name }}] | 21:41 |
| jeblair | 2017-07-07 21:38:44.682497 | TASK [Find tox directories to synchrionize file_type=directory, use_regex=True, patterns=^(?!log).*$, paths=src/{{ zuul.project.canonical_name }}/.tox] | 21:41 |
| mordred | jeblair: both good points :) | 21:41 |
| mordred | jeblair: I think for _these_ roles it's not a super-big issue since it's always going to be about passing in values - but I can totally prefix with zuul_work_dir | 21:42 |
| jeblair | mordred: yeah, i'm still not sure the right answer. | 21:44 |
| jeblair | mordred: as long as we've thought about it and given our best shot given what we know now, i'm happy. :) | 21:45 |
| *** openstackgerrit has quit IRC | 21:47 | |
| mordred | yup | 21:49 |
| *** openstack has joined #zuul | 22:00 | |
| *** openstackgerrit has joined #zuul | 22:02 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add a job that runs zuul's py35 unit tests https://review.openstack.org/481763 | 22:02 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add zuul_work_dir default to one-project roles https://review.openstack.org/481762 | 22:02 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Permit config shadowing https://review.openstack.org/479084 | 22:03 |
| mordred | crap I missed a variable name | 22:04 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add a job that runs zuul's py35 unit tests https://review.openstack.org/481763 | 22:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add zuul_work_dir default to one-project roles https://review.openstack.org/481762 | 22:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Delete run-bindep role https://review.openstack.org/481770 | 22:05 |
| mordred | jeblair: OH - you know what? | 22:12 |
| mordred | jeblair: that's trying to run py35 tests on zuul master | 22:12 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Delete run-bindep role https://review.openstack.org/481770 | 22:13 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add a job that runs zuul's py35 unit tests https://review.openstack.org/481763 | 22:13 |
| mordred | jeblair: I get to play with all the fun toys on this change :) | 22:13 |
| jeblair | mordred: ++ | 22:13 |
| mordred | jeblair: I think I need to port in configure_mirror next | 22:17 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-jobs master: DNM: test per-build ssh key https://review.openstack.org/481772 | 22:17 |
| jeblair | mordred: yeah, i think pabelanger said that was hurting us | 22:17 |
| jeblair | mordred: https://review.openstack.org/473845 | 22:18 |
| mordred | jeblair: cool - I can start with that | 22:20 |
| mordred | jeblair: finger://ze01.openstack.org/ff61c221249d4261a9b7c8d4b79c3b4a ... | 22:22 |
| mordred | jeblair: there's some fun kazoo output up in there that it was blocked on for a while | 22:22 |
| mordred | jeblair: but nevermind | 22:22 |
| jeblair | mordred: i didn't get there in time. which change? | 22:23 |
| mordred | I thought it was zuul output- it was the test output from testing zuul | 22:23 |
| mordred | http://logs.openstack.org/63/481763/5/check/ff61c22/job-output.txt | 22:23 |
| mordred | jeblair: if you scroll to the end of the tests near the bottom you'll see a ton of NFO:paste.httpserver.ThreadPool:Shutting down threadpool | 22:23 |
| jeblair | mordred: yeah, that's normal. the bigger bug there is that we're getting test log output at all in the console. :( | 22:28 |
| mordred | yah | 22:28 |
| jeblair | your stack looks green now, shall i +3 it? | 22:39 |
| jeblair | mordred: ^ | 22:39 |
| mordred | jeblair: yes please | 22:40 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Add zuul_work_dir default to one-project roles https://review.openstack.org/481762 | 22:41 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Delete run-bindep role https://review.openstack.org/481770 | 22:41 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Add a job that runs zuul's py35 unit tests https://review.openstack.org/481763 | 22:41 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Permit config shadowing https://review.openstack.org/479084 | 22:52 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix test_bubblewrap_leak https://review.openstack.org/481123 | 23:12 |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Optionally bind /lib64 https://review.openstack.org/481432 | 23:16 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!