| *** sshnaidm is now known as sshnaidm|off | 00:30 | |
| *** JasonCL has joined #zuul | 00:45 | |
| *** JasonCL has joined #zuul | 00:49 | |
| *** JasonCL has joined #zuul | 00:49 | |
| *** JasonCL has quit IRC | 02:02 | |
| *** JasonCL has joined #zuul | 02:03 | |
| *** JasonCL has joined #zuul | 02:03 | |
| *** JasonCL has quit IRC | 02:33 | |
| *** JasonCL has joined #zuul | 02:39 | |
| *** JasonCL has quit IRC | 02:41 | |
| *** JasonCL has joined #zuul | 02:42 | |
| *** JasonCL has quit IRC | 02:50 | |
| *** JasonCL has joined #zuul | 03:26 | |
| *** bhavik1 has joined #zuul | 03:57 | |
| *** JasonCL has quit IRC | 04:00 | |
| *** bhavik1 has quit IRC | 04:12 | |
| *** bhavik1 has joined #zuul | 04:12 | |
| *** clarkb has quit IRC | 06:02 | |
| *** xinliang has quit IRC | 06:32 | |
| *** clarkb has joined #zuul | 06:33 | |
| *** xinliang has joined #zuul | 06:44 | |
| *** xinliang has quit IRC | 06:44 | |
| *** xinliang has joined #zuul | 06:44 | |
| *** dmellado has quit IRC | 08:06 | |
| *** dmellado has joined #zuul | 08:08 | |
| *** dmellado has quit IRC | 08:18 | |
| *** dmellado has joined #zuul | 08:18 | |
| *** sshnaidm|off is now known as sshnaidm|rover | 08:40 | |
| *** Guest13268 is now known as zigo | 09:09 | |
| *** bhavik1 has quit IRC | 09:27 | |
| *** jaianshu has joined #zuul | 09:30 | |
| jaianshu | Hi | 09:43 |
|---|---|---|
| jaianshu | is this the right forum to ask issues related to nodepool? | 09:44 |
| jaianshu | we are trying to set zuulv3 with nodepool...facing some issue with setup | 09:49 |
| *** hashar has joined #zuul | 10:05 | |
| *** hashar has quit IRC | 10:09 | |
| *** jaianshu has quit IRC | 10:16 | |
| *** jaianshu has joined #zuul | 10:16 | |
| *** jaianshu has quit IRC | 10:23 | |
| *** electrofelix has joined #zuul | 10:24 | |
| *** jaianshu has joined #zuul | 10:28 | |
| *** openstackgerrit has joined #zuul | 11:00 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul feature/zuulv3: web: add /{tenant}/jobs route https://review.openstack.org/503270 | 11:00 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul feature/zuulv3: web: add /{tenant}/builds route https://review.openstack.org/466561 | 11:00 |
| *** jaianshu_ has joined #zuul | 11:02 | |
| *** jaianshu has quit IRC | 11:06 | |
| tristanC | jaianshu_: what issues are you facing? | 11:24 |
| jaianshu_ | tristanC: in my provider i created one nova instance and when i try to get info via 'nodepool alien-list' from host, i don't get any instance info..it shows empty. | 11:28 |
| tristanC | jaianshu_: does it shows when running "OS_CLOUD=provider_name openstack server list" ? | 11:33 |
| *** jesusaur has quit IRC | 11:38 | |
| jaianshu_ | tristanC: yes, it shows here correctly thorough this... | 11:41 |
| jaianshu_ | OS_CLOUD=<> --> cloud name here, not provider name | 11:41 |
| jaianshu_ | i will paste my nodepool.yaml and clould.yaml...1 min | 11:42 |
| tristanC | jaianshu_: i meant when using the cloud attribute of your nodepool provider | 11:43 |
| jaianshu_ | tristanC: yes, getting output from that | 11:44 |
| *** jesusaur has joined #zuul | 11:45 | |
| jaianshu_ | here is my configuration and output : 1) ~/.config/openstack/clouds.yaml: http://paste.openstack.org/show/628069/ 2) | 11:48 |
| jaianshu_ | 3) output of commands: http://paste.openstack.org/show/628071/ | 11:48 |
| jaianshu_ | 2) /etc/nodepool/nodepool.yaml: http://paste.openstack.org/show/628070/ | 11:49 |
| *** jkilpatr has joined #zuul | 11:52 | |
| *** jesusaur has quit IRC | 11:52 | |
| tristanC | jaianshu_: oh, this is nodepoolv2 right? | 11:52 |
| *** jesusaur has joined #zuul | 11:55 | |
| jaianshu_ | tristanC: i installed it feature/zuulv3 branch one | 11:57 |
| *** jkilpatr has quit IRC | 11:57 | |
| jaianshu_ | is this config that i followed belongs to nodepoolv2 ? | 11:57 |
| tristanC | jaianshu_: yes, nodepoolv3 has a slightly different format, see http://git.openstack.org/cgit/openstack-infra/project-config/tree/nodepool/nl01.openstack.org.yaml | 11:59 |
| tristanC | jaianshu_: though, not sure it explains why alien doesn't show up. do you have other issues? | 11:59 |
| jaianshu_ | tristanC: ok..i will update the config and retry once and update | 12:00 |
| *** jesusaur has quit IRC | 12:02 | |
| *** jesusaur has joined #zuul | 12:06 | |
| *** lennyb_ has quit IRC | 12:07 | |
| *** jkilpatr has joined #zuul | 12:10 | |
| Shrews | I am of the opinion now that we should just remove 'alien-list'. It's confusing. I believe it's purpose has morphed from "list leaked instances" (which is handled automatically now by the cleanup thread), to "show me all instances, whether used by nodepool or not". quite by accident, i think | 12:50 |
| Shrews | show me all *unaccounted for* instances | 12:51 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Remove alien_list command https://review.openstack.org/525177 | 12:57 |
| *** jesusaur has quit IRC | 13:00 | |
| *** jesusaur has joined #zuul | 13:02 | |
| *** dkranz has joined #zuul | 13:02 | |
| *** myoung|ruck has quit IRC | 13:35 | |
| *** myoung has joined #zuul | 13:39 | |
| *** jaianshu_ has quit IRC | 13:43 | |
| *** myoung is now known as myoung|ruck | 13:48 | |
| *** jkilpatr has quit IRC | 13:53 | |
| *** dkranz has quit IRC | 14:18 | |
| *** jkilpatr has joined #zuul | 14:23 | |
| *** jkilpatr has quit IRC | 14:35 | |
| *** nguyentrihai has quit IRC | 14:40 | |
| *** nguyentrihai has joined #zuul | 14:40 | |
| *** jkilpatr has joined #zuul | 14:46 | |
| *** jkilpatr has quit IRC | 15:07 | |
| *** dkranz has joined #zuul | 15:25 | |
| Shrews | fyi, having major internet issues today at home. going to relocate to a nearby coffee shop. | 15:26 |
| jeblair | Shrews: agree regarding alien-list | 15:39 |
| jeblair | jianghuaw_: are there any problems other than alien-list? | 15:39 |
| *** jkilpatr has joined #zuul | 15:43 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Remove implied playbook extensions https://review.openstack.org/525239 | 15:44 |
| *** hashar has joined #zuul | 15:46 | |
| jeblair | i have a simple ansible question: can i have an ansible module (implemented in python) use multiple files in the liblary directory? like, i want my module to be able to do something like "from . import utils"... is that possible? | 15:49 |
| *** bhavik1 has joined #zuul | 15:49 | |
| dmsimard | jeblair: question appropriate for #ansible-devel I think. I'm not aware of anyone doing that right now | 15:59 |
| dmsimard | openshift-ansible tends to push these kind of boundaries but I'm not finding anything | 16:00 |
| *** bhavik1 has quit IRC | 16:06 | |
| *** jkilpatr has quit IRC | 16:07 | |
| *** jkilpatr has joined #zuul | 16:07 | |
| mordred | jeblair: there is a structure called module_utils that is what you're looking for | 16:15 |
| jeblair | mordred: thanks! | 16:15 |
| Shrews | jeblair: cool. https://review.openstack.org/525177 removes it | 16:16 |
| mordred | jeblair: https://review.openstack.org/#/c/471921/ is a non-working change that uses the concept | 16:16 |
| mordred | jeblair: that might be useful to as a starting point | 16:16 |
| gundalow | jeblair: Hi. module_utils is where code that's shared between two or more modules should live. Bonus points awarded for added unit tests for such code | 16:17 |
| gundalow | The __init__.py files under lib/ansible/modules/* MUST be empty. | 16:17 |
| mordred | gundalow: jinx ;) | 16:17 |
| gundalow | :D | 16:18 |
| gundalow | Please note the deadlines on https://github.com/ansible/ansible/blob/devel/docs/docsite/rst/roadmap/ROADMAP_2_5.rst | 16:18 |
| gundalow | Core Engine Freeze and Module Freeze: 15 January 2018 | 16:18 |
| gundalow | No more core (which includes module_utils) feature changes after 15th Jan | 16:18 |
| mordred | gundalow: oh, awesome, thanks. I've got a module related PR I need to finish ... | 16:18 |
| gundalow | Ping in #ansible-devel for reviews | 16:19 |
| dmsimard | jeblair, mordred: I think there's a use case where zuul-cloner was convenient and the current implementation of required-projects doesn't work very well for that | 16:20 |
| dmsimard | In TripleO, they currently need to have the whole planet in required-projects which leads the job preparation to take over 5 minutes: https://i.imgur.com/upYA0O1.png | 16:20 |
| dmsimard | wait, I think by typing that out I realized that I'm wrong | 16:21 |
| mordred | dmsimard: :) | 16:21 |
| dmsimard | EmilienM: ^ if you add a project as depends-on, it *will* be prepared, so the zuul-cloner shim will work.. But it'd be worth checking where you are still using zuul-cloner and understanding why we are using it | 16:22 |
| EmilienM | dmsimard: I don't care about git cloning projects, since we use delorean | 16:23 |
| EmilienM | I'm wondering why we git clone actually | 16:23 |
| EmilienM | (in tripleo ecosystem) | 16:23 |
| dmsimard | ok, time to grep zuul-cloner and figure it out :p | 16:24 |
| mordred | EmilienM: for source dependencies, doesn't delorean need to build the packages from something? | 16:28 |
| EmilienM | yeah good question | 16:30 |
| EmilienM | I asked jpena to join here, he's the delorean guru | 16:30 |
| *** jpena has joined #zuul | 16:33 | |
| jpena | hi EmilienM | 16:34 |
| EmilienM | jpena: we're chatting about zuul and delorean - Monty asked: for source dependencies, doesn't delorean need to build the packages from something? | 16:34 |
| EmilienM | jpena: we're wondering if we actually need https://github.com/openstack-infra/tripleo-ci/blob/master/zuul.d/base.yaml#L13 | 16:34 |
| jpena | EmilienM: in the tripleo jobs, dlrn will build using the sources fetched by zuul | 16:36 |
| dmsimard | EmilienM: If the intent is to build a package with the patch that might be in nova if there's a depends-on involved, you need to have that patch checked out somewhere -- otherwise delorean doesn't need you to prepare anything, it's able to retrieve what it needs by itself | 16:36 |
| EmilienM | ok | 16:36 |
| EmilienM | so we need https://github.com/openstack-infra/tripleo-ci/blob/master/zuul.d/base.yaml#L13 I guess | 16:36 |
| EmilienM | unless there is another way | 16:36 |
| EmilienM | to just clone what we need | 16:37 |
| dmsimard | EmilienM: again, you don't, because when there is a Depends-On, there's an "implied" required-projects added | 16:37 |
| dmsimard | EmilienM: it's not added to required-projects but it's checked out by zuul in the list of projects to prepare | 16:37 |
| dmsimard | EmilienM: in ara I could do a depends-on a nova patch and zuul will automatically prepare the repo with that patch | 16:37 |
| EmilienM | cool | 16:37 |
| EmilienM | so let's drop it? | 16:37 |
| dmsimard | EmilienM: the problem lies if you use zuul-cloner, it won't fallback to regular git clone if the project hasn't been prepared first | 16:38 |
| dmsimard | let's move to #tripleo :D | 16:40 |
| *** myoung|ruck is now known as myoung|ruck|bbl | 16:42 | |
| jeblair | dmsimard: if you always want a nova git repo, whether depends-on or not, doesn't that suggest that nova should be in required-projects? | 16:43 |
| *** jpena is now known as jpena|brb | 16:43 | |
| rcarrillocruz | jeblair: heya, is there a tentative date for the feature/zuulv3 merge onto master? | 16:43 |
| dmsimard | jeblair: yeah, but tripleo doesn't always need nova, it only needs it if there is a depends-on | 16:44 |
| jeblair | rcarrillocruz: probably as soon as folks make puppet-openstackci safe for that | 16:44 |
| jeblair | rcarrillocruz: mmedvede is working on it | 16:44 |
| dmsimard | jeblair: but since zuul-cloner is used, it needs to be in required-projects still | 16:44 |
| dmsimard | jeblair: like, in their actual code, they use zuul-cloner | 16:44 |
| dmsimard | (example http://git.openstack.org/cgit/openstack-infra/tripleo-ci/tree/playbooks/tripleo-buildimage/run.yaml#n24 ) | 16:44 |
| jeblair | dmsimard: right, that suggests that it's always "needed" (in the immediate technical sense, if not in the larger abstract sense) | 16:45 |
| dmsimard | correct | 16:45 |
| dmsimard | once we get rid of the zuul-cloner usage and patch holes accordingly, it will no longer necessary to have this larger list of required-projects | 16:45 |
| jeblair | dmsimard: do you want to make it so that jobs either use the prepared repo if it exists, or *no* git repo otherwise? | 16:45 |
| dmsimard | ¯\_(ツ)_/¯ | 16:46 |
| rcarrillocruz | jeblair: ack, thx | 16:46 |
| dmsimard | that was pretty much the behavior of zuul-cloner in v2 | 16:46 |
| jeblair | dmsimard: zuul-cloner in v2 would have always made a git repo. either with a dependent change or not. it should be the same as adding something to required-projects | 16:47 |
| dmsimard | jeblair: right, but what I mean is that right now if the project is not in required-projects, the shim will fail. I'm torn on making the shim "fallback" to a regular clone because, while convenient, it might not incite people to move off of it | 16:47 |
| jeblair | dmsimard: i still don't understand the situation. if you want to mimic the behavior of v2, add the project to required-projects. why doesn't that work here? | 16:49 |
| dmsimard | Might cause bugs around the expectations of what the shim does, like getting a regular git clone when you really wanted a patch checked out | 16:49 |
| dmsimard | jeblair: the behavior of v2 did not force you to declare a list of projects you might need 1% of the time | 16:49 |
| clarkb | dmsimard: it did if you used zuul-cloner though | 16:49 |
| mordred | right. that's more how the base jobs / child jobs are organized | 16:50 |
| jeblair | dmsimard: that is, i think, where we aren't seeing eye-to-eye. if the job *always* uses a git repo, then it requires it. we're just moving that list from within the job itself to the job definition in zuul. in all cases, you still needed/need to, somewhere, say "these are the git repos i need", and for projects like tripleo, it's a long list. | 16:50 |
| dmsimard | clarkb: right but zuul-cloner never failed, either it checked out a patch or it checked out a repo from the tip of the specified branch -- it's also used a bit everywhere in code versus in the job declaration | 16:51 |
| mordred | dmsimard: the same is true of required-projects - either it'll get you a patch or the tip of the branch ... the main shift/difference is using it in job definitions rather than inside of job scripts - it may require rethinking some of the job structure here | 16:53 |
| dmsimard | jeblair: yeah, there are some repos that we always need and where it's appropriate to always have them checked out through required-projects. For other cases (e.g, nova), they're only on a need basis and the code that is where we need to fix code (in tripleo) to make sure zuul-cloner is no longer used. | 16:53 |
| mordred | dmsimard: like, that playbook you pasted should be associated with a job that only lists those 8 repos in its required-projects | 16:54 |
| mordred | or something | 16:55 |
| dmsimard | I think we also have to account for the fact that, unless mistaken, the paths are slightly different between zuul-cloner and required-projects (i.e, workspace/namespace/project vs workspace/src/git.o.o/namespace/project) | 16:55 |
| mordred | right - workspace isn't actually a thing | 16:55 |
| dmsimard | but that's on tripleo to address that | 16:55 |
| jeblair | dmsimard: considering a job that sometimes needs nova: what does that job do with respect to nova in the cases where it doesn't need it? | 16:55 |
| dmsimard | need to brb, picking up kids from school | 16:55 |
| *** jkilpatr has quit IRC | 17:00 | |
| *** kmalloc has joined #zuul | 17:01 | |
| *** nhicher has quit IRC | 17:01 | |
| *** nhicher has joined #zuul | 17:10 | |
| *** nhicher has quit IRC | 17:12 | |
| *** nhicher has joined #zuul | 17:12 | |
| dmsimard | huh so it turns out that they no longer use zuul-cloner in that scenario already | 17:13 |
| dmsimard | they use the git ansible module with a (parsed) refspec to check out | 17:13 |
| dmsimard | I think that's due to constraints when running third party from pure jenkins | 17:14 |
| tobiash | jlk: got the info today that github apps won't make it into ghe 2.12 :( | 17:17 |
| tobiash | jlk: but they're 'actively thinking about it'... | 17:17 |
| tobiash | so hopefully in 2.13 | 17:17 |
| SpamapS | bummer | 17:20 |
| SpamapS | would like to auto-subscribe installers. | 17:20 |
| *** jpena|brb is now known as jpena | 17:25 | |
| *** jkilpatr has joined #zuul | 17:50 | |
| jeblair | dmsimard: i strongly recommend against that approach. preparing repos correctly is difficult, and is one of zuul's main reasons for existing. i'm happy to discuss that further with respect to tripleo in particular in more appropriate channels, but i thought it important to mention here, lest folks get the wrong impression. | 17:58 |
| dmsimard | jeblair: I'm not in disagreement, however it is an approach that works if you're running from outside Zuul and want to retrieve that depends-on | 17:59 |
| dmsimard | all ears if there's another approach we should be taking | 17:59 |
| jeblair | dmsimard: let's choose a channel other than #zuul to discuss running things outside of zuul | 18:01 |
| dmsimard | #openstack-infra or #tripleo ? | 18:01 |
| jeblair | either of those sound better, yes | 18:02 |
| *** hashar is now known as hasharAway | 18:05 | |
| *** jpena is now known as jpena|away | 18:20 | |
| *** sshnaidm|rover is now known as sshnaidm|off | 18:23 | |
| *** jpena|away is now known as jpena|off | 18:48 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul feature/zuulv3: WIP: Add finger gateway https://review.openstack.org/525276 | 18:58 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add command socket support to zuul-merger https://review.openstack.org/523197 | 19:17 |
| *** myoung|ruck|bbl is now known as myoung|ruck | 19:21 | |
| *** jkilpatr has quit IRC | 19:34 | |
| *** jkilpatr has joined #zuul | 21:33 | |
| dmsimard | jeblair: is there better documentation on how to use "start" somewhere ? https://docs.openstack.org/infra/zuul/feature/zuulv3/user/config.html#attr-pipeline.start | 21:35 |
| openstackgerrit | Sean McGinnis proposed openstack-infra/zuul-jobs master: Fix variable name in npm-publish job https://review.openstack.org/525326 | 21:36 |
| *** dkranz has quit IRC | 21:37 | |
| dmsimard | jeblair: I guess my real question is -- I'm trying to see if I can get zuul to post a message when it starts jobs in another pipeline a bit like how we do with gate jobs: http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul.d/pipelines.yaml#n87 | 21:38 |
| jeblair | dmsimard: yeah, the detailed docs are driver dependent; i don't think we've figured out how to cross-ref them there | 21:43 |
| jeblair | dmsimard: so look under "reporter configuration" here for the driver that applies: https://docs.openstack.org/infra/zuul/feature/zuulv3/admin/connections.html | 21:44 |
| pabelanger | I think it is start_message, but cannot find it in docs | 21:45 |
| jeblair | hrm. the docs for the gerrit reporter for leave something to be desired. they should be expanded to look like the ones for github. | 21:45 |
| *** hasharAway has quit IRC | 21:46 | |
| pabelanger | yah, source says start-message, but looks like we don't document it | 21:46 |
| dmsimard | jeblair: an idea along those lines that I had a while back was this: https://storyboard.openstack.org/#!/story/2000873 | 21:53 |
| dmsimard | but I know that your opinion is to keep zuul as hidden as possible :) | 21:53 |
| dmsimard | Meeting in 5 minutes in #openstack-meeting-alt right ? | 21:55 |
| * dmsimard fixed his calendar | 21:55 | |
| jeblair | dmsimard: i think zuul has support for that already. it was reverted from hideci.js in openstack because no one, including its author, wanted to expend the effort to fix hideci.js. | 21:55 |
| jeblair | dmsimard: but the github driver uses it for its status link | 21:56 |
| jeblair | should still work just as well for gerrit | 21:56 |
| dmsimard | interesting, didn't know | 21:56 |
| jeblair | dmsimard: be aware gerrit-next has a dedicated channel for ci reporting, which changes all this substantially. | 21:56 |
| dmsimard | neat | 21:57 |
| jeblair | dmsimard: re meeting: yes | 21:58 |
| jeblair | meeting now in #openstack-meeting-alt | 22:00 |
| dmsimard | mordred, pabelanger, Shrews, fungi, clarkb ^ | 22:01 |
| dmsimard | SpamapS: ^ | 22:01 |
| *** myoung|ruck is now known as myoung|ruck|bbl | 22:11 | |
| *** myoung|ruck|bbl is now known as myoung|ruck | 22:11 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Move send_command() into ZuulApp https://review.openstack.org/523211 | 22:14 |
| *** adam_g has quit IRC | 22:22 | |
| *** adam_g has joined #zuul | 22:24 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
