| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: WIP: Add logo and move menu to top right https://review.openstack.org/565909 | 00:45 |
|---|---|---|
| corvus | clarkb: ^ that's with better padding (for all sizes) and a stuck another menu item in there 'cause i'm sure we'll have at least one more, so it'll give us a better idea | 00:46 |
| corvus | clarkb: if it still look too out-there on extra-wide screens, we can add more padding | 00:46 |
| pabelanger | corvus: clarkb: I just noticed on ^ i no longer see the list of jobs zuul ran under code-review votes on right side of page | 00:52 |
| pabelanger | moved the discussion to #openstack-infra | 00:54 |
| *** threestrands has joined #zuul | 01:05 | |
| *** Guest10014 is now known as jlvillal | 01:19 | |
| *** spsurya has joined #zuul | 02:54 | |
| openstackgerrit | Merged openstack-infra/zuul master: Update documentation URLs https://review.openstack.org/565910 | 04:11 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Don't create db accounts unencrypted https://review.openstack.org/565948 | 04:15 |
| *** jpena|off has quit IRC | 04:29 | |
| *** ssbarnea has quit IRC | 04:29 | |
| *** weshay has quit IRC | 04:29 | |
| *** fbo_ has quit IRC | 04:30 | |
| *** fbo_ has joined #zuul | 04:31 | |
| *** mhu` has quit IRC | 04:31 | |
| *** jpena|off has joined #zuul | 04:31 | |
| *** pabelanger has quit IRC | 04:31 | |
| *** nhicher has quit IRC | 04:31 | |
| *** ssbarnea has joined #zuul | 04:32 | |
| *** pabelanger has joined #zuul | 04:32 | |
| *** mhu has joined #zuul | 04:32 | |
| *** nhicher has joined #zuul | 04:32 | |
| *** weshay has joined #zuul | 04:35 | |
| SpamapS | hrm.. so.. I have my tenants setup to pull in the git repos for our local mirrors of openstack repos.. | 05:07 |
| SpamapS | and I have them set to ignore project/project-template/job | 05:07 |
| SpamapS | but now I have one complaining about a required project in a job in one of those repos | 05:08 |
| clarkb | SpamapS: meaning its not properly ignoring that config from the upstream repo? | 05:09 |
| clarkb | SpamapS: if you updated in the last few days the big config rewrite may be to blame. We've had a few bugs in it that we've been fixing as we find them | 05:09 |
| SpamapS | Yeah I think I found one | 05:10 |
| SpamapS | just updated recently but I may be a few commits behind | 05:11 |
| clarkb | it did great things for our memory consumption and we don't have to trickle config changes in which is nice | 05:11 |
| SpamapS | I should probably setup a gate job on my local zuul mirror to see if all my configs still load ;) | 05:11 |
| clarkb | but was major refactoring and htere has been some fallout | 05:11 |
| SpamapS | k | 05:11 |
| SpamapS | let me first just try the obvious step of comparing what's been merged since I deployed Monday | 05:12 |
| SpamapS | (Deployed Monday, but just now restarting scheduler) | 05:12 |
| SpamapS | I guess I deployed yesterday actually | 05:13 |
| SpamapS | 8c88420b044059cd977684638bf8e4c090f98e1b | 05:13 |
| SpamapS | and since then.... | 05:13 |
| SpamapS | well.. nothing obvious | 05:14 |
| SpamapS | but I'll go ahead and kick a deploy to bring in the latest | 05:14 |
| clarkb | SpamapS: the latest may not fix that specific bug its not familiar to me but should fix one or two others | 05:14 |
| SpamapS | indeed... I'm dropping those from the tenant file as the jobs that use them are non-voting and can be skipped for a while. | 05:17 |
| SpamapS | Will hopefully be able to debug on my stage env and figure out what's broken | 05:19 |
| SpamapS | I have a lot of repos listed that exclude stuff actually | 05:20 |
| SpamapS | because I use zuul as much as a "grab all the things form git and gather them" as I do for in-line configs. | 05:20 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Add tox-py36 jobs https://review.openstack.org/565881 | 05:31 |
| tobiash | pabelanger: added a depends-on to a postgres db fix on bionic ^ | 05:33 |
| SpamapS | reproduced on master | 05:37 |
| SpamapS | looks like excluded jobs with required-projects still try to find the required projects | 05:37 |
| SpamapS | Hm | 05:46 |
| SpamapS | this is sort of nasty | 05:46 |
| SpamapS | so it looks like we always still *parse* everything | 05:46 |
| SpamapS | even if we're going to exclude stuff | 05:46 |
| SpamapS | so any validation done in the parser (fromYaml) will still have to pass | 05:46 |
| SpamapS | Instead of using the exclude to decide not to parse | 05:47 |
| SpamapS | we are using it to decide not to load | 05:47 |
| openstackgerrit | Merged openstack-infra/zuul master: Don't create db accounts unencrypted https://review.openstack.org/565948 | 06:18 |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul master: Do not fail so early on missing required-projects https://review.openstack.org/565973 | 06:19 |
| SpamapS | ^^ that might fix it | 06:19 |
| SpamapS | might break other stuff tho | 06:20 |
| SpamapS | woot tests pass | 06:32 |
| SpamapS | clarkb: ^ I believe that fixes the bug... might want to prioritize it in case anybody else is trying to pull in mirrors. | 06:33 |
| *** zhuli has quit IRC | 06:50 | |
| tobiash | SpamapS: I don't see the place where this validation is done later, maybe you want to add a negative test case with a non-existent required-project | 06:53 |
| *** threestrands has quit IRC | 06:58 | |
| *** gtema has joined #zuul | 07:01 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: WIP: test required project not found error https://review.openstack.org/565979 | 07:07 |
| tobiash | SpamapS: that should illustrate what i mean ^ | 07:08 |
| *** sshnaidm|rover has quit IRC | 07:08 | |
| *** ssbarnea_ has joined #zuul | 07:30 | |
| *** ssbarnea_ has quit IRC | 07:31 | |
| *** jpena|off is now known as jpena | 07:35 | |
| *** ssbarnea_ has joined #zuul | 07:45 | |
| *** sshnaidm|rover has joined #zuul | 07:55 | |
| *** electrofelix has joined #zuul | 07:57 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Don't validate references for excluded config items https://review.openstack.org/565973 | 07:58 |
| tobiash | SpamapS, corvus, clarkb: added a further test case and using a different approach for the fix ^ | 07:59 |
| tobiash | passes local test suite | 08:00 |
| tobiash | SpamapS: hoping it's ok that I hijacked your change | 08:00 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Only emit parent-change-enqueued if needed https://review.openstack.org/563194 | 08:48 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Add tox-py36 job https://review.openstack.org/565884 | 09:03 |
| openstackgerrit | Merged openstack-infra/zuul master: Trivial: Update pypi url to new url https://review.openstack.org/563248 | 09:27 |
| SpamapS | tobiash: no, thank you. I was having trouble wrapping my head around how to do it the way you did it. :) | 09:31 |
| SpamapS | Stared at the code for a long, long time and then just decided that validation was more trouble than it was worth. ;) | 09:31 |
| tobiash | SpamapS: well I just had the advantage that I reviewed the config refactor stack | 09:39 |
| *** yolanda_ has joined #zuul | 09:49 | |
| *** yolanda has quit IRC | 09:51 | |
| cmurphy | how does pull request gating in github work with zuul? i guess you'd have to delegate merging to a bot instead of pushing the merge button? | 09:55 |
| *** yolanda has joined #zuul | 09:56 | |
| tobiash | cmurphy: yes, zuul basically does the same as with gerrit (pressing the merge button via api) | 09:56 |
| *** yolanda_ has quit IRC | 09:59 | |
| *** yolanda has quit IRC | 10:02 | |
| rcarrillocruz | cmurphy: fbo_ wrote this nice post about it https://blogs.rdoproject.org/2017/12/gate-repositories-on-github-with-software-factory-and-zuul3/ | 10:09 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Extend github testing using app auth https://review.openstack.org/563297 | 10:12 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Test parent-change-enqueued with github https://review.openstack.org/563242 | 10:12 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Optionally record created github clients in tests https://review.openstack.org/563402 | 10:12 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Limit search scope of getChangesDependingOn to tenant https://review.openstack.org/563196 | 10:13 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Remove getPullBySha from FakeGithubConnection https://review.openstack.org/563416 | 10:13 |
| *** yolanda has joined #zuul | 10:15 | |
| cmurphy | rcarrillocruz: thanks for that link, there's not a good example like that in the docs | 10:15 |
| *** yolanda has quit IRC | 10:25 | |
| *** yolanda has joined #zuul | 10:39 | |
| *** jpena is now known as jpena|lunch | 10:58 | |
| *** persia has joined #zuul | 11:06 | |
| sshnaidm|rover | dmsimard, is it possible to add another dashboard to grafana for tripleo-ci? something like "CI status" | 11:35 |
| *** sshnaidm|rover is now known as sshnaidm|rover|a | 12:08 | |
| *** sshnaidm|rover|a is now known as sshnaidm|rover | 12:09 | |
| *** jpena|lunch is now known as jpena | 12:11 | |
| *** dkranz has joined #zuul | 12:18 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul master: Improve test output on assertHistory fail https://review.openstack.org/566048 | 12:28 |
| tobiash | This hopefully can help us to track down random failures like http://logs.openstack.org/16/563416/5/check/tox-py35/7c56097/testr_results.html.gz ^ | 12:28 |
| *** rlandy has joined #zuul | 12:34 | |
| openstackgerrit | Fatih Degirmenci proposed openstack-infra/zuul master: Add CentOS 7 environment setup instructions https://review.openstack.org/564948 | 12:39 |
| *** yolanda has quit IRC | 12:57 | |
| *** yolanda has joined #zuul | 13:11 | |
| *** ssbarnea_ has quit IRC | 13:59 | |
| *** acozine has joined #zuul | 14:11 | |
| *** yolanda has quit IRC | 14:19 | |
| *** sshnaidm|rover is now known as sshnaidm|afk | 14:26 | |
| *** pabelanger has quit IRC | 14:35 | |
| *** pabelanger has joined #zuul | 14:35 | |
| pabelanger | tobiash: thanks! | 14:35 |
| corvus | tobiash, pabelanger: search for "Build history:" in the log output there | 14:46 |
| tobiash | corvus: thanks, abandoned | 14:48 |
| pabelanger | I'm sorry, search where? | 14:49 |
| tobiash | pabelanger: there: http://logs.openstack.org/16/563416/5/check/tox-py35/7c56097/testr_results.html.gz | 14:49 |
| tobiash | corvus: the question is why is one build aborted in this test case where we expect all builds with success... | 14:50 |
| *** yolanda has joined #zuul | 14:53 | |
| gundalow | Hi all. I'm creating a new image for use with Zuul/nodepool, are their any details on the requirements, such as users and public keys that I must bake into the base image? | 15:17 |
| gundalow | Or the steps/source for building your existing Fedora images would be great | 15:26 |
| Shrews | corvus: i found a snag for being able to support multiple labels per node. deals with stats reporting: http://git.openstack.org/cgit/openstack-infra/nodepool/tree/nodepool/stats.py#n110 | 15:28 |
| clarkb | gundalow: zuul has to be able to ssh in (or use winrm I think that is the windows connection system) You can bake the keys into the image to make this happen or use some boot time configuration system like cloud-init to set it up | 15:32 |
| clarkb | gundalow: we bake it in, our nodepool configs for our images can be found at https://git.openstack.org/cgit/openstack-infra/project-config/tree/nodepool/nodepool.yaml#n103 | 15:32 |
| clarkb | gundalow: nodepool-base depends on zuul-worker which set s that up for us https://git.openstack.org/cgit/openstack-infra/project-config/tree/nodepool/elements/zuul-worker useradd and sudo stuff at https://git.openstack.org/cgit/openstack-infra/project-config/tree/nodepool/elements/zuul-worker/install.d/60-zuul-worker | 15:33 |
| corvus | Shrews: hrm. good question. we could count each label independently, i guesss. that math will come out a bit funny. or we could stop reporting label stats. or we could drop back to just one label per node. | 15:35 |
| corvus | is anyone here using the static driver? | 15:36 |
| Shrews | corvus: counting each independently would probably make that stat useless, yeah? | 15:36 |
| corvus | mrhillsman, tristanC, SpamapS, tobiash ^ | 15:36 |
| corvus | Shrews: i guess it depends on how you're using it | 15:37 |
| mrhillsman | reading | 15:37 |
| clarkb | tobiash: SpamapS I'll review the config fix post tea ingestion | 15:37 |
| corvus | mrhillsman: sorry, just was trying to ask if you're using the static driver | 15:37 |
| mrhillsman | nah | 15:37 |
| mrhillsman | not yet, only openstack right now | 15:37 |
| corvus | Shrews: maybe "i have 10 bigmem nodes ready" is a useful stat, even though we'd also be saying "i have 5 bigcpu nodes ready" and those 5 are also included in the 10 | 15:38 |
| mrhillsman | with the email that went out the other day about additional drivers i wanted to start investigating additional drivers but will probably not happen until after vancouver | 15:39 |
| corvus | Shrews: obviously if you use one bigcpu, both counters will decrement by one, so adding up your total nodes ready by adding up the labels would be a bad idea. but you could still do it by provider and get an accurate number. | 15:40 |
| corvus | Shrews: anyway, i think we should ask folks using the static driver if we really need to support multiple labels. a *lot* of stuff gets easier if we don't. | 15:40 |
| Shrews | corvus: maybe i should start a ML discussion | 15:40 |
| corvus | Shrews: sounds good | 15:41 |
| Shrews | i do think supporting multiple breaks a LOT of assumptions we've made | 15:41 |
| *** ssbarnea_ has joined #zuul | 15:43 | |
| gundalow | clarkb: Ace, thank you | 15:43 |
| corvus | clarkb: how does http://logs.openstack.org/09/565909/2/check/zuul-website-build/d2fc189/html/ look on your uss enterprise bridge viewscreen? | 15:47 |
| tobiash | corvus: we're using the static driver | 15:48 |
| clarkb | corvus: thats better. its still in the right corner but now with a lot of padding | 15:48 |
| tobiash | but didn't know that it is possible to have multiple labels per static node | 15:49 |
| corvus | tobiash: great -- here's a question: are you using multiple labels for the same nodes? | 15:49 |
| corvus | tobiash: perfect. pretend it is not possible. :) | 15:49 |
| tobiash | :) | 15:49 |
| corvus | tobiash: do you think you may end up needing something like that? | 15:50 |
| tobiash | I don't think so currently | 15:50 |
| corvus | tobiash: ok, thanks. it's one of those things where it's easy to come up with a theoretical use for, but it turns out it makes a lot of things hard, so we're trying to get an idea of how important it is, or might be. | 15:51 |
| tobiash | it could make sharing static nodes a bit easier but tbh it sounds awkward | 15:51 |
| clarkb | I think if you come from the jenkins idea of a label you amy want it | 15:51 |
| clarkb | since labels there are attribute tags then your job describes I want x86 with 8GBmemory and the fastssd disk and ubuntu xenial and 10Gbe Nics | 15:52 |
| clarkb | and if a single node happens to meet all those criteria its what you get otherwise the job doesn't schedule or queues iirc | 15:52 |
| clarkb | whereas we've used it as a high level this node looks like this other node with a single label | 15:53 |
| corvus | clarkb: yeah, i'm also kind of worried about us ending up with something like the jenkins "(label || label) && label" thing. i'm not sure if supporting multiple labels on the nodepool side makes us more likely to end up with that or less. :/ | 15:54 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: WIP: Add fedora mirror to nodepool dsvm jobs https://review.openstack.org/566102 | 15:55 |
| clarkb | I have tea now. I feel like all code review requires tea | 15:55 |
| tobiash | corvus, clarkb our multi label would probably be the opposite of jenkins | 15:56 |
| mordred | I think also the primary design and use of zuul/nodepool is dynamic nodes - while static nodes are a bit more like "I want the machine that's connected to the cargo bay door controller" | 16:00 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Add logo and move menu to top right https://review.openstack.org/565909 | 16:01 |
| clarkb | SpamapS: tobiash ok config exclusion loading fix should be on its way in | 16:01 |
| mordred | the use cases for really complex flexible use of resources are really better served by cloud - and I'd be worried that we'd be implementing a de-facto cloud control plane in nodepool if we try to make the static driver *too* powerful | 16:01 |
| corvus | mordred: that's a good point, and i love the example and think that's what we should put in the docs. | 16:01 |
| mordred | corvus: yay! I'm helpful | 16:01 |
| corvus | clarkb, SpamapS, tobiash: bonus: that fix has a small net memory savings :) | 16:02 |
| clarkb | related to reviewing that change, I'm really happy with where the test suite has ended up. Helps review changes so much. however if I get a free afternoon I think we could probably stand to do some house keeping in there. test_scheduler is basically test everything and test_v3 is test everything else. | 16:03 |
| corvus | clarkb: that is exactly right. | 16:03 |
| clarkb | (and now that v3 is behind us test_v3 just seems wrong, its implied) | 16:03 |
| corvus | clarkb: yes, but we should do something other than "move everything into test_scheduler" so that'll take at least a little thought :) | 16:04 |
| clarkb | ++ | 16:05 |
| corvus | clarkb: 565909 is unwip now; i removed the community menu item since it was just a placeholder; i will add it back in a later change | 16:06 |
| mordred | clarkb: we could move it all into test_scheduler and then rename thatto "test_everything" | 16:06 |
| corvus | mordred: i feel like you have brilliantly solved the problem! | 16:06 |
| *** yolanda has quit IRC | 16:06 | |
| openstackgerrit | Merged openstack-infra/zuul master: Don't validate references for excluded config items https://review.openstack.org/565973 | 16:11 |
| Shrews | ML email sent | 16:14 |
| Shrews | and clarkb has cleverly planted the desire for tea in my head, so i will go make that now | 16:14 |
| openstackgerrit | Merged openstack-infra/zuul-website master: Replace background image https://review.openstack.org/565908 | 16:15 |
| clarkb | Shrews: my problem now is I have run out | 16:15 |
| Shrews | mmm, rasberry zinger | 16:20 |
| clarkb | corvus: mordred I think there is room there to group by function under test. Ya the scheduler is involved in most of it but many of those tests are parser or executor etc focused and we can group around that? | 16:22 |
| corvus | clarkb: yes, i think it's worth trying to group in file by function like that. note that in the new system, it's easy to have tests which use the same configuration in the same class, so further clustering them like that makes sense | 16:24 |
| SpamapS | corvus: oh yay | 16:27 |
| SpamapS | mordred: are you saying that when we have static working right, I can write "Zuul. Open the pod bay doors." | 16:29 |
| mordred | SpamapS: yes | 16:29 |
| SpamapS | "I'm sorry, I can't do that Clint. You need write access to this git repository." | 16:29 |
| clarkb | did you see they are doing a new 70mm run? | 16:29 |
| mordred | SpamapS: it is your job now to make the "Zuul. Open the pod bay doors." shirt | 16:29 |
| clarkb | from restored originals? | 16:30 |
| mordred | clarkb: WHAT? NO! I must to see that | 16:30 |
| clarkb | mordred: ya I don't recall all details but you should definitely look it up if you are interested | 16:30 |
| Shrews | I have just come up with the best zuul tagline: "Because humans suck" | 16:31 |
| clarkb | Rerelease opens in select US cities starting may 18 | 16:32 |
| mordred | Shrews: :) | 16:35 |
| corvus | clarkb, mordred: *unrestored* :) | 16:42 |
| corvus | they just struck new 70mm prints :) | 16:43 |
| clarkb | ah cool. Also neat that after 50 years they are able to do that considering how difficult its been for some films | 16:44 |
| corvus | mordred: it should be going to the alamo drafthouse in austin in july | 16:50 |
| corvus | http://www.in70mm.com/news/2018/one_movie/index.htm | 16:51 |
| clarkb | oh wow we are getting it at the hollywood | 16:52 |
| *** yolanda has joined #zuul | 16:53 | |
| corvus | i was hoping it would be in vancouver; but the only ca showing i've found is toronto in several months. | 17:00 |
| *** yolanda_ has joined #zuul | 17:01 | |
| *** jpena is now known as jpena|off | 17:01 | |
| *** yolanda has quit IRC | 17:04 | |
| *** yolanda has joined #zuul | 17:05 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: WIP: Add fedora mirror to nodepool dsvm jobs https://review.openstack.org/566102 | 17:05 |
| *** yolanda_ has quit IRC | 17:08 | |
| *** ssbarnea_ has quit IRC | 17:14 | |
| *** gtema has quit IRC | 17:20 | |
| *** spsurya has quit IRC | 17:42 | |
| clarkb | pabelanger: comment on https://review.openstack.org/#/c/565881/5 that will make the script work on more than just ubuntu | 17:56 |
| clarkb | pabelanger: not a -1 though but may be a nice improvment | 17:56 |
| pabelanger | clarkb: yah, good idea | 17:58 |
| clarkb | that said I think on suse at least apt-get update && apt-get install bubblewrap would work (because zypper is magic) | 18:00 |
| clarkb | corvus: did you see tristanC's note about executable parameter to the command module? | 18:15 |
| clarkb | corvus: we may want to do a scripted update to all the openstack configs nowish to update that which ahs the added beneift of further testing the config changes | 18:15 |
| mordred | clarkb: where is that note - I missed it | 18:16 |
| clarkb | asking you about it since you've done the large scripted updates in the past | 18:16 |
| clarkb | mordred: one sec I will get you a link | 18:16 |
| clarkb | mordred: http://lists.zuul-ci.org/pipermail/zuul-discuss/2018-May/000194.html and https://review.openstack.org/#/c/546869/ | 18:16 |
| *** yolanda has quit IRC | 18:20 | |
| mordred | feels like turning that to shell: type -p bindep args: executable: /bin/bash intead of command: /bin/bash -c "type -p bindep" would be a better route if we were to do a global scripted update | 18:21 |
| pabelanger | anybody else see: openstack.exceptions.ConfigException: Cloud fake was not found. from nodepool master? | 18:22 |
| clarkb | mordred: I think the reason for that is shell uses sh by default and type -p may not be bourne | 18:22 |
| mordred | for task in tasks: if 'command' in task and task.get('args', {}).get('executable'): task['shell'] = task.pop('command') | 18:22 |
| clarkb | mordred: but yes we can address that using shell | 18:22 |
| mordred | clarkb: right - but shell does still accept executable | 18:22 |
| clarkb | mordred: ya | 18:22 |
| pabelanger | http://logs.openstack.org/02/566102/2/check/tox-py35/d02b614/testr_results.html.gz | 18:22 |
| pabelanger | seems os-client-config was just released | 18:23 |
| pabelanger | wonder if that broke us | 18:23 |
| mordred | pabelanger: ah - if it did, we should fix it | 18:25 |
| pabelanger | downgrading to 1.30.0 seems to fix it | 18:25 |
| pabelanger | looking to see what the diff is | 18:25 |
| mordred | *excellent* | 18:25 |
| mordred | well - the diff is that 1.31 rips out the guts | 18:25 |
| Shrews | yeah, that was a biggish change | 18:25 |
| * mordred starts poking | 18:26 | |
| pabelanger | oh | 18:26 |
| clarkb | heh locally I get tox complaining about testr not being a valid command to setup.py | 18:27 |
| clarkb | maybe we should get the switch to stestr in | 18:27 |
| mordred | ok. I reproducedit locally | 18:28 |
| Shrews | clarkb: https://review.openstack.org/536862 seems tobiash didn't +A it for some reason. feel free to do so | 18:28 |
| pabelanger | mordred: yay | 18:28 |
| clarkb | Shrews: thanks will look | 18:29 |
| clarkb | tox -re py36 fixed my immediate issue though | 18:29 |
| mordred | ah- ok, I think I see what's wrong | 18:31 |
| clarkb | setUpFakes monkey patches oscc | 18:34 |
| mordred | yup | 18:34 |
| clarkb | which is no longer going to be a working path with the move. Should be an easy fix thankfully | 18:34 |
| mordred | yah- the larger question is - should we figure out a fix in the os-client-config shim layer to allow this code to work as written? | 18:34 |
| clarkb | if it was outside of test code I would say yes | 18:35 |
| clarkb | considering its monkeypatching in test code this is probably the risk we assume for doing that? | 18:35 |
| mordred | the OpenStackConfig object does take a list of config files as an argument which is how I would normally expect people to interact with that | 18:35 |
| *** yolanda has joined #zuul | 18:35 | |
| clarkb | ya I think the reason its patched like this is we don't expose that config to nodepool users | 18:36 |
| clarkb | they are expected to use the normal paths? | 18:36 |
| clarkb | (maybe we should expose it and then just set it in the tests though | 18:36 |
| mordred | yah | 18:36 |
| clarkb | mordred: probably for now just update the monkeypatch then we can add in config for that if we think it would be useful | 18:38 |
| mordred | ok. I hve a fix for os-client-config | 18:41 |
| clarkb | mordred: + 'openstack.config.loader.CONFIG_FILES', [clouds_path])) works on the nodepool side | 18:44 |
| mordred | clarkb: great. let's do that for now - I've got a fix that fixes the current usage - but I believe it will break that ^^ | 18:45 |
| mordred | because yay | 18:45 |
| clarkb | ok I'll push that up | 18:45 |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool master: Fix test patching of clouds.yaml file locations https://review.openstack.org/566138 | 18:47 |
| clarkb | and no sense in approving stestr change until ^ is merged I guess | 18:47 |
| mordred | clarkb: ok, I think this remote: https://review.openstack.org/566139 Allow monkeypatching CONFIG_FILES | 18:48 |
| mordred | clarkb: should fix it in both directions | 18:48 |
| mordred | nope | 18:48 |
| mordred | only in the one still | 18:48 |
| clarkb | out of curiousity why import loader and import * from laoder? | 18:49 |
| clarkb | (noticed then when figuring out where CONFIG_FILES is defined) | 18:49 |
| mordred | not sure the first import loader is actually necessary | 18:50 |
| clarkb | it is for line 22 but that could be updated | 18:50 |
| clarkb | its just weird to do both | 18:50 |
| mordred | clarkb: remote: https://review.openstack.org/566139 Allow monkeypatching CONFIG_FILES | 18:53 |
| mordred | now works with both directions of monkeypatch | 18:53 |
| mordred | clarkb: oh - I remember now | 18:54 |
| mordred | get_one_cloud = loader.OpenStackConfig.get_one | 18:54 |
| mordred | if we don't import loader we don't have a handle to that for the method alias | 18:55 |
| clarkb | mordred: that actually makes me question the import * more beacuse os_client_config.CONFIG_FILES is openstack.config.loader.CONFIG_FILES after that import | 18:55 |
| clarkb | then if you monkey patch only one will be patched | 18:55 |
| clarkb | whcih may lead to confusion though maybe not in this case | 18:55 |
| mordred | yup - that's why the most recent patch had to do copy of the original | 18:56 |
| mordred | to compare against | 18:56 |
| clarkb | mordred: right why not stop importing * and just scope everything to loader? I'm guessing the original intent was to make this overrideable except it isn't due to this? though maybe thats specific to passing in args | 18:56 |
| clarkb | ya I guess that is why it does import * its so those names show up properly | 18:57 |
| mordred | yah | 18:57 |
| mordred | a person who was just doing "from os_client_config.config import CONFIG_FILES" should not get broken - otherwise we'd have to do explicit CONFIG_FILES = loader.CONFIG_FILES for each of the globals which just seemed ick ... and still doens't fix the monkeypath issue :) | 18:57 |
| clarkb | ya | 18:58 |
| mordred | pabelanger: https://review.openstack.org/#/c/566138 should fix the tests for you | 18:58 |
| mordred | pabelanger: and then https://review.openstack.org/566139 should fix os-client-config in case anyone else has that problem | 18:58 |
| mordred | Shrews: ^^ | 18:58 |
| pabelanger | great, will look shortly | 18:59 |
| *** hashar has joined #zuul | 19:04 | |
| *** openstackgerrit has quit IRC | 19:05 | |
| *** ssbarnea_ has joined #zuul | 19:29 | |
| *** gtema has joined #zuul | 19:49 | |
| *** openstackgerrit has joined #zuul | 19:50 | |
| openstackgerrit | Clark Boylan proposed openstack-infra/nodepool master: Fix test patching of clouds.yaml file locations https://review.openstack.org/566138 | 19:50 |
| clarkb | turns out there was more than one location to update ^ | 19:50 |
| clarkb | also we should do a followup to switch to openstacksdk in the deps and openstack driver | 19:51 |
| *** ssbarnea_ has quit IRC | 19:54 | |
| *** ssbarnea_ has joined #zuul | 19:55 | |
| *** gtema has quit IRC | 20:10 | |
| mordred | clarkb: yah. I've been meaning to make some patches for that | 20:25 |
| *** weshay is now known as weshay|rover | 20:30 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool master: Use openstacksdk instead of os-client-config https://review.openstack.org/566158 | 20:50 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool master: Remove use of six https://review.openstack.org/566159 | 20:50 |
| mordred | corvus: I'm not 100% sure about the exception re-raise code in the remove use of six patch | 20:51 |
| *** dkranz has quit IRC | 20:54 | |
| mordred | corvus: nevermind - http://paste.openstack.org/show/720324/ - that seems to verify that the code used in the patch should be correct | 20:56 |
| corvus | mordred: yay chicken! | 21:01 |
| *** acozine has quit IRC | 21:06 | |
| *** dtruong_ has quit IRC | 21:13 | |
| corvus | looks like there's some suggestions for mission statements and taglines in https://etherpad.openstack.org/p/zuul-mission | 21:45 |
| corvus | friendly reminder to get suggestions in there soon, we'll pick one next week | 21:45 |
| *** hashar has quit IRC | 22:08 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Add community page https://review.openstack.org/566179 | 22:17 |
| *** rlandy is now known as rlandy|bbl | 22:30 | |
| *** threestrands has joined #zuul | 22:59 | |
| *** gouthamr has quit IRC | 23:08 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Switch to Open Sans font https://review.openstack.org/566185 | 23:10 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul-website master: Change tagline to dark blue https://review.openstack.org/566186 | 23:10 |
| *** gouthamr has joined #zuul | 23:10 | |
| *** gouthamr has quit IRC | 23:18 | |
| *** dmellado has quit IRC | 23:19 | |
| SpamapS | Is there an infra/zuul dinner planned for Vancouver? | 23:44 |
| * SpamapS may have missed emails | 23:44 | |
| clarkb | SpamapS: I sent email to the infra list collecting names and availability it is getting close enoguh I should figure out what we are doing | 23:48 |
| clarkb | looks like monday night is best so far | 23:48 |
| *** gouthamr has joined #zuul | 23:51 | |
| *** dmellado has joined #zuul | 23:51 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!