| -openstackstatus- NOTICE: Zuul has been restarted to update to the latest code; existing changes have been re-enqueued, you may need to recheck changes uploaded in the past 10 minutes | 00:00 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul master: Update github payload location docs https://review.openstack.org/557550 | 00:15 |
|---|---|---|
| corvus | pabelanger, tobiash, mordred: ^ that should be fixed before release. | 00:16 |
| pabelanger | +2 | 00:16 |
| corvus | nodepool 3.0.0 tagged | 00:27 |
| corvus | https://pypi.python.org/pypi/nodepool/3.0.0 exists | 00:31 |
| openstackgerrit | Merged openstack-infra/zuul master: Update github payload location docs https://review.openstack.org/557550 | 00:45 |
| corvus | zuul 3.0.0 tagged | 00:46 |
| pabelanger | Yay, congrats all | 00:47 |
| jlk | WOO | 00:50 |
| clarkb | now to find a drink once landed | 00:50 |
| pabelanger | I'm doing that myself now | 00:51 |
| *** odyssey4me has quit IRC | 00:53 | |
| *** odyssey4me has joined #zuul | 00:53 | |
| corvus | https://pypi.python.org/pypi/zuul/3.0.0 exists | 00:54 |
| corvus | i think that means we actually released this thing :) | 00:54 |
| corvus | i just sent an announcement to zuul-announce | 00:54 |
| pabelanger | ++ | 00:55 |
| *** ChanServ sets mode: -c | 00:57 | |
| jlk | hey wow, guess I"ll have a bit of whiskey tonight! | 00:57 |
| corvus | indeed! | 00:57 |
| jlk | brought home some Irish whiskey, seems like a good reason | 00:57 |
| corvus | let's make it festive in here | 00:57 |
| *** ChanServ changes topic to "Zuul 3.0.0 release party | Website: https://zuul-ci.org/ | Docs: https://zuul-ci.org/docs/ | Source: https://git.zuul-ci.org/ | Channel logs: http://eavesdrop.openstack.org/irclogs/%23zuul/" | 00:58 | |
| pabelanger | Yah, I see a Zuul 3.0.0 released email | 00:59 |
| corvus | Zuul v3 is released! | 00:59 |
| clarkb | I think I'm 30k feet over corvus' house too | 00:59 |
| corvus | clarkb: cheers! | 00:59 |
| clarkb | corvus: indeed! | 01:00 |
| corvus | okay, i'm going to go out and eat and drink, and will continue celebrating with everyone tomorrow! | 01:00 |
| pabelanger | well deserved, if I don't say so myself | 01:01 |
| corvus | thanks everyone! | 01:01 |
| *** JasonCL has quit IRC | 01:03 | |
| *** JasonCL has joined #zuul | 01:04 | |
| *** JasonCL has joined #zuul | 01:09 | |
| *** JasonCL has quit IRC | 01:11 | |
| *** JasonCL has joined #zuul | 01:12 | |
| *** JasonCL has quit IRC | 01:13 | |
| *** JasonCL has joined #zuul | 01:13 | |
| * fungi cheers | 01:18 | |
| *** JasonCL has quit IRC | 01:49 | |
| *** JasonCL has joined #zuul | 01:49 | |
| *** JasonCL has quit IRC | 01:54 | |
| *** JasonCL has joined #zuul | 01:55 | |
| *** JasonCL has quit IRC | 02:13 | |
| mrhillsman | woo hoo! | 02:23 |
| mrhillsman | quite late to the party but i made it :) | 02:23 |
| EmilienM | congrats folks, can't wait for zuul v4 ;-) | 02:28 |
| mrhillsman | lol | 02:30 |
| mrhillsman | soooooo | 02:30 |
| mrhillsman | do not shoot darts/arrows/etc my way but i have a question/request/plead | 02:30 |
| mrhillsman | we have done quite a bit of work with k8s as some may be aware re openlab (gophercloud/terraform/cloud-provider-openstack) | 02:32 |
| dmsimard | congrats everyone :D | 02:32 |
| mrhillsman | openlab using zuul of course | 02:32 |
| mrhillsman | we requested to add github app into kubernetes/cloud-provider-openstack annnnnnnnd...not going to happen as of right now | 02:33 |
| mrhillsman | suggested approach is to add a driver/plugin/connector for prow+zuul | 02:33 |
| mrhillsman | there is some stuff that was done for jenkins which was mentioned as potential reference | 02:34 |
| mrhillsman | https://github.com/kubernetes/test-infra/tree/master/prow/jenkins | 02:34 |
| mrhillsman | needless to say, it is all go | 02:34 |
| mrhillsman | so asking for some eyes/minds to provide some suggestions other than having to develop a driver/plugin/connector | 02:35 |
| clarkb | mrhillsman: it would probably help if some of the thought behind that was provided too | 02:35 |
| clarkb | so that we can understand the need | 02:36 |
| *** dims_ has joined #zuul | 02:36 | |
| mrhillsman | dims should be joining in a sec to help with that | 02:37 |
| clarkb | (and actually address that rather than making a random plugin because someone said "no") | 02:37 |
| dims_ | o/ | 02:37 |
| mrhillsman | sure | 02:37 |
| mrhillsman | hence my asking for thoughts/eyes on this of you all who know the things better than me | 02:38 |
| dims_ | http://paste.openstack.org/show/716958/ | 02:38 |
| mrhillsman | dims_ can you see the stuff i typed earlier? | 02:38 |
| mrhillsman | ok cool, was not sure if that was shareable | 02:39 |
| dims_ | public slack ... | 02:39 |
| dims_ | just was not sure if everyone here had accounts | 02:40 |
| corvus | so, if i understand correctly, you're offering your services as a third-party ci system for a kubernetes repo. and they have rejected that offer because they don't like the technology you're running. | 02:40 |
| corvus | and that would be akin to openstack requiring all third-party ci systems to run zuul. | 02:40 |
| dims_ | corvus : it's not a rejection, exploring options | 02:40 |
| corvus | in fact, we don't do that, because that's not very friendly. we have plenty of folks running jenkins third-party ci systems reporting on openstack. | 02:41 |
| tristanC | yay a tag, finally :-) | 02:41 |
| corvus | dims_: oh, okay, i misunderstood mrhillsman's "annnnnnnnd...not going to happen as of right now" | 02:41 |
| mrhillsman | hehe, yes, yes you did | 02:41 |
| corvus | so what is the situation? | 02:42 |
| mrhillsman | sorry, i should have been more tactful | 02:42 |
| dims_ | "This would be better because a) other prow users would then have access to OpenLab CI and/or Zuul and b) it will interface better with the rest of our automation like tide (so /retest would work, for example)" <<< very valid! | 02:42 |
| clarkb | a) would just be a matter if them adding the app too | 02:42 |
| corvus | what's "/retest" ? | 02:43 |
| clarkb | not sure thats an actual concern | 02:43 |
| dims_ | corvus : prow has a jenkins plugin. so they asked if we could look into a similar one for zuul | 02:46 |
| dims_ | issue is here - https://github.com/kubernetes/test-infra/issues/7371 | 02:46 |
| pabelanger | if you keep your job config out of kubernetes/cloud-provider-openstack, you don't need to add the bot | 02:48 |
| pabelanger | we are actually testing kubernetes-incubator/kubespray with zuul.o.o today | 02:48 |
| corvus | pabelanger: how so? i suspect they'd like to report on pull requests. | 02:48 |
| dims_ | pabelanger : can we do testing and reporting on PR(s)? | 02:48 |
| dims_ | yes corvus | 02:48 |
| mrhillsman | ^ | 02:49 |
| pabelanger | right, we don't report on PRs | 02:49 |
| dims_ | pabelanger : we need that | 02:49 |
| pabelanger | dims_: right, but test-infra basically said no right now | 02:49 |
| dims_ | pabelanger : they asked me to check on something and i am here trying to do that | 02:50 |
| mrhillsman | dims mentioned also maybe write a zuul driver based on timer that uses github v3 notifications api | 02:50 |
| pabelanger | but agree, it is pretty limiting to only say prowl for testing | 02:50 |
| corvus | what's "/retest"? | 02:50 |
| dims_ | pabelanger : we can do periodic right away and report results to the central dashboard | 02:50 |
| dims_ | they do want us to do that for sure | 02:51 |
| mrhillsman | idk corvus - thought dims_ would be able to answer that | 02:51 |
| mrhillsman | i suspect it is the same as recheck | 02:51 |
| *** JasonCL has joined #zuul | 02:52 | |
| dims_ | it's one of the bot commands | 02:52 |
| mrhillsman | found it | 02:52 |
| mrhillsman | https://prow.k8s.io/command-help | 02:52 |
| dims_ | yes like recheck | 02:52 |
| mrhillsman | Rerun test jobs that have failed. | 02:52 |
| corvus | if so, then it's probably a very small change to the pipeline config in openlab to support it; of course it's mrhillsman's decision on whether that's desirable. but in general, i'd say that if there's a concern about matching their environment -- well, zuul is *very* configurable in that regard. | 02:53 |
| corvus | to put it succinctly, it's going to be way easier to add 2 lines of yaml to support "/retest" than it would be to write plugins | 02:53 |
| corvus | dims_, mrhillsman: i'll try to speculate on more direct answers to your question, but i have limited knowledge of prow | 02:54 |
| dims_ | corvus : the main thing is how do we push events from prow to zuul so we could do pr(s) | 02:55 |
| corvus | in general, zuul has a driver interface for supporting multiple trigger sources. it may be possible to create one that is compatible in some way with prow. i'm not sure what form that would take -- is it push? pull? it'll take some thought. | 02:55 |
| corvus | similarly, reporters are a driver interface as well. so it might be possible to report to prow. | 02:55 |
| dims_ | either we write a prow plugin to send something from prow to zuul ... or just use github notifications api as the source of events to zuul | 02:56 |
| corvus | dims_: well, the github notification api is "install the openlab github app" | 02:56 |
| pabelanger | is there a passive way to get some events with installing github app? | 02:57 |
| dims_ | not that one ... "GET /notifications" (from https://developer.github.com/v3/activity/notifications/) then they don't have to install github app or configure our webhook | 02:57 |
| pabelanger | without* | 02:58 |
| dims_ | y pabelanger ^^ | 02:58 |
| corvus | dims_: what's the rate limit on that? | 02:59 |
| dims_ | pabelanger : zuul has a timer driver right, we can use that with GET /notifications?since= | 02:59 |
| dims_ | corvus : dunno will have to check | 02:59 |
| clarkb | corvus: I think all github api is 5k requests per hour | 02:59 |
| mrhillsman | it would be per user yes? | 02:59 |
| pabelanger | yah, I guess that is the down side of not adding the app, lower rate limit? | 03:00 |
| corvus | mrhillsman: something like openlab could eat that *very* quickly. | 03:00 |
| corvus | it takes a lot of api requests to do anything with github | 03:00 |
| mrhillsman | ah ok | 03:01 |
| *** JasonCL has quit IRC | 03:01 | |
| dims_ | create a new "bot" for each repo that we would like to be notified about? | 03:01 |
| mrhillsman | i thought it was a push rather than a pull | 03:01 |
| corvus | dims_: i think that the notification api is not going to be a good experience for users (due to the lag and the impact on requests). there is a significant benefit to using app installs. that's what github encourages folks like us to do. | 03:01 |
| mrhillsman | while github app is a push | 03:01 |
| dims_ | ack corvus . will point @fejta to this discussion | 03:02 |
| dims_ | corvus : the other one was "if prow can talk to jenkins api, can we not write one that talks to zuul api" | 03:03 |
| dims_ | that would be a push from prow plugin calling something in zuul | 03:03 |
| corvus | dims_: zuul doesn't have a generalized trigger api, yet, though it potentially could. that's a better solution than the notifications api, however, it's still going to need to talk to github to learn about the actual PRs, so we'll run into the rate limit too. | 03:04 |
| corvus | (for that matter, i'd argue jenkins doesn't really have one either -- that's why we wrote the gearman plugin) | 03:05 |
| corvus | (it sort of has a rest api for triggering jobs, but it's gotten more and more limited over time) | 03:05 |
| dims_ | ack thanks for the chat i think i have enough to take back to them and keep the conversation going | 03:07 |
| dims_ | thanks pabelanger corvus clarkb | 03:07 |
| mrhillsman | ^ | 03:07 |
| corvus | dims_: np. keep me posted. :) | 03:07 |
| mrhillsman | sorry to interrupt the party :) | 03:07 |
| dims_ | will do | 03:08 |
| corvus | dims_, mrhillsman: you *must* have a drink before you go. :) | 03:08 |
| pabelanger | dims_: reading the backscroll, I didn't see a specific reason not to allow the openlab app into k8s. Might be good to better understand why that isn't liked | 03:08 |
| * mrhillsman sips twice | 03:08 | |
| dims_ | ah apologies. congrats on the new release! | 03:08 |
| dims_ | where are my manners | 03:08 |
| mrhillsman | i don't think it was not liked pabelanger rather than are there other options | 03:08 |
| * mrhillsman puts words in bad order at times causing confusion | 03:09 | |
| dims_ | pabelanger : that's not ruled out yet, just looking for better ux for folks who already use prow infra | 03:09 |
| pabelanger | ack | 03:09 |
| corvus | yeah, very valid concern -- we have the same in openstack -- we just proscribe the UX and are ambivalent about the tooling | 03:09 |
| corvus | dims_, mrhillsman: cheers! | 03:10 |
| mrhillsman | thank you! | 03:10 |
| tristanC | corvus: fwiw, https://review.openstack.org/554839 demonstrate a POST api/job/$job-name to trigger a single job | 03:23 |
| *** harlowja has quit IRC | 03:24 | |
| tristanC | this even creates a little web form with the job variables :-) | 03:24 |
| *** JasonCL has joined #zuul | 03:25 | |
| *** JasonCL has quit IRC | 03:29 | |
| openstackgerrit | Bernd Bausch proposed openstack-infra/zuul master: Add clarity to *Zuul from Scratch* instructions https://review.openstack.org/556344 | 03:30 |
| *** patriciadomin has quit IRC | 03:50 | |
| *** patriciadomin has joined #zuul | 03:51 | |
| *** bhavik1 has joined #zuul | 03:57 | |
| *** harlowja has joined #zuul | 03:59 | |
| *** jimi_|ansible is now known as jimi|ansible | 04:12 | |
| jimi|ansible | congrats all :) | 04:13 |
| * jimi|ansible is late to the party | 04:13 | |
| tobiash | Yay release :) | 04:20 |
| openstackgerrit | Bernd Bausch proposed openstack-infra/zuul master: Add clarity to *Zuul from Scratch* instructions https://review.openstack.org/556344 | 04:57 |
| openstackgerrit | Bernd Bausch proposed openstack-infra/zuul master: Add clarity to *Zuul from Scratch* instructions https://review.openstack.org/556344 | 05:27 |
| openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul master: Add clarity to *Zuul from Scratch* instructions https://review.openstack.org/556344 | 05:29 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: add gentoo systemd image for testing https://review.openstack.org/556661 | 05:37 |
| openstackgerrit | Ian Wienand proposed openstack-infra/nodepool master: Test growroot in boot tests https://review.openstack.org/555103 | 05:37 |
| *** harlowja has quit IRC | 06:06 | |
| *** flepied__ has quit IRC | 06:19 | |
| openstackgerrit | Merged openstack-infra/nodepool master: add gentoo systemd image for testing https://review.openstack.org/556661 | 06:31 |
| *** flepied__ has joined #zuul | 06:53 | |
| *** JasonCL has joined #zuul | 07:43 | |
| *** JasonCL has quit IRC | 07:48 | |
| *** electrofelix has joined #zuul | 07:54 | |
| *** flepied_ has joined #zuul | 08:00 | |
| *** flepied__ has quit IRC | 08:01 | |
| *** flepied__ has joined #zuul | 08:02 | |
| *** flepied_ has quit IRC | 08:04 | |
| *** flepied_ has joined #zuul | 08:08 | |
| *** flepied__ has quit IRC | 08:10 | |
| *** flepied__ has joined #zuul | 08:14 | |
| *** flepied_ has quit IRC | 08:16 | |
| *** hashar has joined #zuul | 08:21 | |
| *** flepied__ has quit IRC | 08:32 | |
| *** sshnaidm is now known as sshnaidm|afk | 08:33 | |
| *** JasonCL has joined #zuul | 08:52 | |
| *** JasonCL has quit IRC | 08:57 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul master: zuul-changes: update for the new api url https://review.openstack.org/557656 | 09:13 |
| *** JasonCL has joined #zuul | 10:09 | |
| *** sshnaidm|afk is now known as sshnaidm | 10:23 | |
| *** flepied has joined #zuul | 10:48 | |
| *** bhavik1 has quit IRC | 10:51 | |
| *** JasonCL has quit IRC | 11:00 | |
| tobiash | jlk, corvus: found another problem with the github3.py: http://paste.openstack.org/show/717856/ | 11:27 |
| tobiash | :( | 11:27 |
| tobiash | I don't know yet in which cases this occurs | 11:29 |
| tobiash | but frequently inhibits changes to be enqueued | 11:29 |
| tristanC | tobiash: it doesn't seems to happen with public github.com | 11:29 |
| tobiash | tristanC: yah, that's only a problem with github enterprise | 11:29 |
| electrofelix | we have a few people asking for us to support squashing instead of merging, while I might disagree with the idea, imo should consider the commit state as part of the review | 11:45 |
| electrofelix | but it looks like something that will have to be done for those using github locally | 11:45 |
| electrofelix | what approach would be acceptable? make the most sense to deliver upstream? | 11:45 |
| electrofelix | in repo configuration option to 'merge' or 'squash' on submit? | 11:47 |
| electrofelix | support enabling of the 'squash' option only in the github repo conf and query for configured behaviour (if possible) or simply try 'squash' if 'merge' receives an error saying not allowed? | 11:47 |
| electrofelix | or support a second label controlling whether to 'merge' or 'squash' and perform the action depending on which label is applied? | 11:47 |
| tobiash | electrofelix: you mean using githubs functionality merge with sqash? | 11:59 |
| tobiash | we probably should extend the github driver that it can select the desired merge method | 11:59 |
| tobiash | there is already a per repo setting of a merge mode | 12:00 |
| tobiash | but I don't know if that only applies to merging in zuul or if it's also intended to select the merge method of gerrit/github | 12:00 |
| electrofelix | tobiash: I couldn't see that any of the code attempting to perform a squash instead of merge? | 12:04 |
| electrofelix | mergePull at https://github.com/openstack-infra/zuul/blob/master/zuul/driver/github/githubconnection.py#L1065 seems to be the only method called | 12:04 |
| tobiash | electrofelix: yes, that's unimplemented currently | 12:04 |
| electrofelix | yeah, that's what I thought. we're still on zuulv2 so I was hoping to get an idea of what 'squash merge PR' might be in zuulv3 in the future so I can try and retrofit something close enough in the mean time | 12:05 |
| tobiash | the github3.py supports the merge method: http://github3.readthedocs.io/en/latest/pulls.html?highlight=merge#github3.pulls.PullRequest.merge | 12:06 |
| electrofelix | tobiash: right, I guess I'm really trying to get a feel for what method of exposing that to github projects might be acceptable in zuulv3 to see if I can mimic it in our v2 so there is no change in behaviour or need to carry patches when we do migreate | 12:14 |
| tobiash | electrofelix: I think this or a submit-mode sibling would do the job: https://zuul-ci.org/docs/zuul/user/config.html#attr-project.merge-mode | 12:17 |
| tobiash | so we either could link that to the driver or add a submit-mode setting to that | 12:17 |
| *** elyezer has quit IRC | 12:33 | |
| *** elyezer has joined #zuul | 12:34 | |
| electrofelix | tobiash: cool, so in repo control only? | 12:34 |
| electrofelix | s/control/config/ | 12:34 |
| tobiash | yes, that's the idea | 12:34 |
| tobiash | you already now can configure the merge behavior in zuul itself | 12:34 |
| tobiash | so I think it makes sense to make the submit method configurable the same way | 12:35 |
| tobiash | tristanC: just found out the github3.py problem is not limited to github enterprise | 12:42 |
| tobiash | it occurs when fetching the files from a pr and you added a large file (>3000 lines) | 12:43 |
| tobiash | then github skips the patch field | 12:43 |
| tobiash | like curl -s https://api.github.com/repos/tobiashenkel/sandbox/pulls/1/files | jq . | 12:43 |
| electrofelix | tobiash: thanks, I guess for zuulv2 it'll have to be a hack with either using the configured settings/retrying or support a custom label | 12:45 |
| electrofelix | tobiash: does github require a particular header to return large files? e.g. by default the web interface doesn't display large diffs either | 12:46 |
| *** electrofelix has quit IRC | 12:46 | |
| tobiash | I don't think so it just doesn't include it and you need to get the content via the content links | 12:47 |
| tobiash | also this behavior doesn't seem to be documented: https://developer.github.com/v3/pulls/#list-pull-requests-files | 12:48 |
| tobiash | corvus, jlk: https://github.com/sigmavirus24/github3.py/issues/816 | 12:49 |
| tobiash | tristanC: ^ | 12:49 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: Reorganzie "Zuul From Scratch" document https://review.openstack.org/556988 | 12:53 |
| *** odyssey4me has quit IRC | 12:57 | |
| *** odyssey4me has joined #zuul | 12:57 | |
| Shrews | wow, i cannot spell | 12:59 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/zuul master: Reorganize "Zuul From Scratch" document https://review.openstack.org/556988 | 12:59 |
| tobiash | any gearman cracks here? | 13:00 |
| tobiash | I have a frozen zuul-web waiting for gearman response after a scheduler restart | 13:01 |
| pabelanger | yah, believe you still need to restart it manually | 13:02 |
| tobiash | that doesn't occur every time | 13:02 |
| tobiash | just sometimes | 13:02 |
| tobiash | normally it just reconnects | 13:02 |
| tobiash | http://paste.openstack.org/show/717897/ | 13:03 |
| pabelanger | when we stop / start scheduler on zuul.o.o we also have to do zuul-web too | 13:03 |
| pabelanger | tobiash: is scheduler running gearman server? | 13:04 |
| tobiash | yes | 13:04 |
| pabelanger | work around would be to use geard, but yah we need to fix it | 13:04 |
| tobiash | ah, ok so I'll work around that with liveness probes for now :) | 13:05 |
| pabelanger | yah, believe so. we might be able to reproduce with unit test too | 13:08 |
| pabelanger | tobiash: github question, are you using exclude-unprotected-branches setting at all? https://zuul-ci.org/docs/zuul/admin/tenants.html#attr-tenant.exclude-unprotected-branches | 13:08 |
| tobiash | pabelanger: I use that for all github repos | 13:08 |
| tobiash | it's crucion if you don't work with forking | 13:09 |
| pabelanger | kk, that's what I figured | 13:09 |
| pabelanger | curious why it is disabled by default | 13:09 |
| tobiash | in github enterprise the more common working model is just branching and pr | 13:09 |
| tobiash | it's disabled by default because of the common work pattern fork&pull on github.com | 13:10 |
| pabelanger | ah, so this will not allow that? | 13:11 |
| tobiash | I'm not sure if that really makes sense but I'm ok with that and just set it on every tenant | 13:11 |
| tobiash | with fork&pull it's not really an issue because the speculative source is in a different repo (which is not known by zuul) | 13:12 |
| pabelanger | yah | 13:12 |
| tobiash | but as soon as you allow branch&pull it's an issue | 13:12 |
| pabelanger | yes, that is what happening now. branches are getting created on project zuul is on, then PR from it, which is breaking zuul config | 13:12 |
| pabelanger | As I understood, branch protection and this setting help protect that | 13:13 |
| tobiash | yes | 13:13 |
| pabelanger | ++ | 13:13 |
| pabelanger | thank you | 13:13 |
| tobiash | enable branch protection and this setting and zuul will ignore any config in a non protected branch | 13:13 |
| tobiash | you even can set this per repo btw | 13:13 |
| pabelanger | yah, I think we can do that to test | 13:14 |
| *** flepied_ has joined #zuul | 13:28 | |
| *** flepied has quit IRC | 13:31 | |
| *** JasonCL has joined #zuul | 13:48 | |
| *** JasonCL has quit IRC | 13:52 | |
| *** JasonCL has joined #zuul | 13:52 | |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Fix ansible-lint job https://review.openstack.org/557507 | 13:52 |
| *** JasonCL has quit IRC | 13:55 | |
| *** JasonCL has joined #zuul | 13:56 | |
| *** JasonCL has quit IRC | 13:58 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Upgrade from angularjs (v1) to angular (v5) https://review.openstack.org/551989 | 13:59 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Use glyphicons for status balls https://review.openstack.org/551992 | 13:59 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Stop falling back to job name for missing url https://review.openstack.org/554056 | 13:59 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Remove dashboard workaround for missing log_url https://review.openstack.org/554066 | 13:59 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Add zuul-build-dashboard-multi-tenant https://review.openstack.org/557713 | 13:59 |
| *** flepied__ has joined #zuul | 14:07 | |
| *** JasonCL has joined #zuul | 14:07 | |
| *** flepied_ has quit IRC | 14:09 | |
| *** dkranz has joined #zuul | 14:12 | |
| *** JasonCL has quit IRC | 14:12 | |
| *** JasonCL has joined #zuul | 14:13 | |
| dmsimard | mordred: the new glyphicons look nice! | 14:15 |
| *** jlvillal has quit IRC | 14:15 | |
| dmsimard | it looks like some of them could be pushed up just a tiny bit (2-3 pixels?), triggered my OCD https://i.imgur.com/M0J6yLT.png | 14:17 |
| *** JasonCL has quit IRC | 14:24 | |
| Shrews | Am I so far removed from web dev that I've never heard the term "glyphicon" before? I think I'm happy about that | 14:28 |
| *** JasonCL has joined #zuul | 14:29 | |
| *** flepied_ has joined #zuul | 14:30 | |
| *** maeca has joined #zuul | 14:33 | |
| *** flepied__ has quit IRC | 14:33 | |
| *** JasonCL has quit IRC | 14:35 | |
| mordred | Shrews: :) | 14:37 |
| mordred | Shrews: it's the web-dev version of windings - a font that contains a set of symbols you can use for things without having to use image files | 14:38 |
| mordred | Shrews: I put up a patch suggesting we use them instead of the balls on the status page: http://logs.openstack.org/92/551992/11/check/zuul-build-dashboard/373e86e/npm/html/status.html | 14:40 |
| pabelanger | tobiash: mind a review on https://review.openstack.org/556624/ since we talked about it the other day | 14:42 |
| pabelanger | Shrews: may you also ^ | 14:42 |
| Shrews | pabelanger: do host keys ever get populated in our tests? | 14:52 |
| Shrews | oh yes, i see the test above expects them | 14:53 |
| pabelanger | yah | 14:53 |
| Shrews | pabelanger: 1 minor suggestion | 14:56 |
| Shrews | otherwise lgtm | 14:57 |
| *** JasonCL has joined #zuul | 14:57 | |
| *** bramwelt has quit IRC | 14:58 | |
| pabelanger | k | 14:58 |
| Shrews | pabelanger: oh! ignore that | 15:00 |
| Shrews | pabelanger: i didn't realize that it was already under the OpenStack driver section | 15:00 |
| pabelanger | np | 15:00 |
| Shrews | +2'd | 15:01 |
| *** JasonCL has quit IRC | 15:01 | |
| *** JasonCL has joined #zuul | 15:02 | |
| *** JasonCL has quit IRC | 15:04 | |
| pabelanger | tyty | 15:04 |
| *** JasonCL has joined #zuul | 15:04 | |
| SpamapS | ^5's to everyone for the 3.0.0 release! :-D | 15:06 |
| *** JasonCL has quit IRC | 15:06 | |
| * SpamapS goes back into the salt mine | 15:06 | |
| *** JasonCL has joined #zuul | 15:08 | |
| *** JasonCL has quit IRC | 15:11 | |
| *** JasonCL has joined #zuul | 15:11 | |
| *** jlvillal has joined #zuul | 15:12 | |
| Shrews | How do folks feel about splitting up the "Zuul From Scratch" doc into separate docs for different distros/configurations, as I've suggested here: http://logs.openstack.org/88/556988/3/check/build-sphinx-docs/23aea68/html/admin/zuul-from-scratch.html | 15:13 |
| Shrews | I'd like to begin adding docs for a static nodepool, and a gerrit setup, and possibly some ubuntu specific setup | 15:13 |
| Shrews | alternative, we could just do an entirely separate document | 15:14 |
| Shrews | but there is enough in common, this felt like the better way | 15:14 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: Update pypi metadata https://review.openstack.org/557738 | 15:14 |
| pabelanger | Shrews: corvus: is ^ the correct metadata now? Not that I'm trying to remove the accomplishments of the openstack infra team | 15:16 |
| corvus | Shrews: openstack was unable to muster the resources to keep all of the installation methods up to date | 15:17 |
| corvus | (i mean documentation of all the methods) | 15:18 |
| *** JasonCL has quit IRC | 15:18 | |
| corvus | i worry that as soon as we add a second, the first will be out of date. i'm not saying we should always stick with fedora27 (after all, it'll be out of date automatically in a few months?) | 15:19 |
| Shrews | caveat emptor | 15:19 |
| corvus | but i wonder whether having a fedora and ubuntu section is sustainable | 15:19 |
| corvus | (especially since the differences are not significant) | 15:20 |
| Shrews | corvus: i'm less enthusiastic about the distro part. it's the nodepool driver and repo backends i'd like to see some docs for | 15:20 |
| corvus | (though now that i look at it, what, we only have like 10 lines of fedora-specific stuff in there? maybe i'm over thinking this :) | 15:21 |
| corvus | Shrews: oh, maybe i misunderstood your question | 15:21 |
| Shrews | If someone wants to try zuul, but doesn't have access to an openstack cloud, showing how to setup the static driver with some VMs could be valuable. | 15:21 |
| corvus | Shrews: i wholeheartedly agree there. i definitely think we should add that | 15:22 |
| Shrews | And there was LOTS I learned about gerrit trying to set it up myself (stuff we don't have documented) | 15:22 |
| Shrews | just need someplace to vomit my learnings | 15:22 |
| Shrews | I can leave the fedora stuff as-is | 15:23 |
| corvus | Shrews: i think i'd like to walk back what i said about fedora/ubuntu; cause i think i like your change as is | 15:23 |
| corvus | Shrews: and that has me wondering if we should move everything in zuul-from-scratch.rst into its own pages, so you're only looking at one section at a time. | 15:27 |
| corvus | Shrews: (ie, move "zuul and nodepool installation" into a page, etc) | 15:27 |
| Shrews | corvus: i'm open to whatever seems best. That might make the entire doc easier to follow | 15:29 |
| corvus | Shrews: i'm +2 on the current change, we can continue down that path in a followup if you want | 15:29 |
| Shrews | i noticed some open reviews from others for that doc, so maybe we should land those first and then i can rework mine? | 15:30 |
| corvus | Shrews: probably a good idea -- https://review.openstack.org/556344 ? | 15:30 |
| Shrews | corvus: yeah | 15:31 |
| corvus | Shrews: i'm reviewing that now | 15:31 |
| Shrews | "For experimentation, DevStack is fine" --- LOL, good luck with that if you try to run it in a separate vm | 15:34 |
| Shrews | i wasted so many hours trying to get that to work | 15:34 |
| clarkb | corvus: Shrews re install docs my personal opinion on that is don't specify a distro at all. Install zookeeper from upstream tarball and zuul to a virtualenv. Bindep will mostly let us hand wave around the distro differences in this case I think | 15:35 |
| Shrews | clarkb: yeah, i'd prefer if we had never assumed a particular distro for that | 15:36 |
| corvus | clarkb: yeah, we're basically there already. http://logs.openstack.org/88/556988/3/check/build-sphinx-docs/23aea68/html/admin/fedora27_setup.html | 15:36 |
| corvus | clarkb: it's only the top 3 blockquotes which are distro specific | 15:36 |
| clarkb | maybe assume systemd so that you can copy the zk init scripts to the right location and systemctl daemon-reload. But that should be mostly safe at this point as most distros are using it (even gentoo has an option for systemd) | 15:37 |
| Shrews | i think we can split the service management stuff out too | 15:37 |
| corvus | clarkb: fedora and ubuntu provide zookeeper packages though, i think it's friendlier to new users to tell them to use those. | 15:38 |
| corvus | clarkb: yeah, we do use systemd | 15:38 |
| corvus | Shrews: and yes, we can probably carry those files in zuul somewhere; that will reduce the bulk of the instructions nicely | 15:38 |
| Shrews | corvus: ooh, that would be nice | 15:38 |
| corvus | clarkb: maybe we can re-order the bindep stuff so that the "sudo dnf install git redhat-lsb-core python3 python3-pip python3-devel make gcc openssl-devel python-openstackclient -y" step isn't necessary. | 15:39 |
| clarkb | corvus: ya bindep just needs a python install (2 or 3) and lsb_release maybe we can specify those are dependencies then go from there assuming bindep? | 15:40 |
| pabelanger | +1 | 15:40 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Upgrade to webpack 4 https://review.openstack.org/551987 | 15:41 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Upgrade from angularjs (v1) to angular (v5) https://review.openstack.org/551989 | 15:41 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul master: Use glyphicons for status balls https://review.openstack.org/551992 | 15:41 |
| corvus | clarkb: i bet "dnf install python lsb_release" or "apt install python lsb_release" won't be too hard for us to maintain | 15:43 |
| corvus | once we get that down to the minimum | 15:43 |
| corvus | Shrews, mrhillsman: i left questions on https://review.openstack.org/556344 -- perhaps as recent users of that, you can offer insight | 15:44 |
| corvus | clarkb: ^ | 15:44 |
| Shrews | corvus: yeah saw those. i do not have answers though :( | 15:45 |
| Shrews | and i actually never "used" that doc since i did something TOTALLY different (ran from git repos and used gerrit) | 15:46 |
| mrhillsman | sure thing corvus | 15:46 |
| *** JasonCL has joined #zuul | 15:47 | |
| mrhillsman | as a general blanket statement i take zuul from scratch as the thing i should use to deploy something i can expect to use in production | 15:47 |
| Shrews | mrhillsman: you would recommend an all-in-one for production? | 15:47 |
| mrhillsman | and getting started as a quick start allinone just want to try out and get up and running | 15:47 |
| corvus | Shrews: aio for prod should be fine | 15:48 |
| corvus | it's a matter of scale, not sustainability | 15:48 |
| corvus | mrhillsman: i think the thing trying to be conveyed by the name was that it would walk you through the whole process, without the crutches of any pre-made images, etc. | 15:48 |
| mrhillsman | and getting started as a quick start just want to try out and get up and running | 15:49 |
| mrhillsman | yes | 15:49 |
| clarkb | corvus: ++ one of my biggest gripes with the move to automagic container images is you never learn how anything works (and consequently have little idea of where to look when things break) | 15:49 |
| mrhillsman | and of course it can grow as noted by the disclaimer at the top | 15:49 |
| mrhillsman | just a quick skim of the review | 15:50 |
| corvus | clarkb: yeah, a thing leifmadsen says is that, at some point, you need the documentation that the person making the container images uses to figure out what they need to do | 15:50 |
| mrhillsman | i think some of the things there are because of some pieces being too specific | 15:50 |
| mrhillsman | i.e. useradd vs adduser | 15:50 |
| *** JasonCL has quit IRC | 15:50 | |
| corvus | i don't understand that one -- i'm pretty sure i copy/pasted that when i ran through; is adduser wrong? | 15:52 |
| mrhillsman | i don't know what bernd used | 15:52 |
| mrhillsman | but i used ubuntu | 15:52 |
| mrhillsman | and adduser was/is not available | 15:52 |
| mrhillsman | i did not try to install it because i knew to use useradd :) | 15:53 |
| corvus | mrhillsman: it's available in the adduser package on ubuntu | 15:53 |
| mrhillsman | unfortunately i have not got to spend another iteration of zfs | 15:54 |
| Shrews | i have both available on my ubuntu | 15:54 |
| mrhillsman | but an example i have in the notes i think will help me explain better | 15:54 |
| mrhillsman | https://etherpad.openstack.org/p/zuul-quickstart-and-zfs(zuul-from-scratch) | 15:54 |
| mrhillsman | under the Section: Environment Setup | 15:54 |
| corvus | but that's a good thing to know. perhaps bernd also used ubuntu. the instructions are currently for fedora. i'm in favor of normalizing the instructions to support both, but we need to be clear about what we're doing. we can't just switch between the two and end up with a document that says "use fedora and run this ubuntu command" | 15:54 |
| *** JasonCL has joined #zuul | 15:54 | |
| mrhillsman | like the package names for fedora of course are different than ubuntu and some others i am sure | 15:54 |
| mrhillsman | so if you look at that etherpad | 15:55 |
| mrhillsman | i hoped to go through sections and do stuff like that section | 15:55 |
| mrhillsman | more generic info even though fedora is used | 15:56 |
| mrhillsman | windows and macos may never be supported/able to be used, but i was basically trying to make that statement | 15:57 |
| mrhillsman | you need these packages, however you get them, you should know | 15:57 |
| corvus | mrhillsman: yeah, we discussed minimizing ways to minimize the packages section earlier (15:38) | 15:58 |
| mrhillsman | zuul and nodepool install pieces could explain why you need the users and the home dirs where they are vs /home/$USER | 15:59 |
| mrhillsman | i say that based on doing install without these instructions and getting perms errors from the executor | 15:59 |
| corvus | mrhillsman: you're going to push up changes at some point, right? :) | 15:59 |
| mrhillsman | yes sir | 15:59 |
| mrhillsman | what i will do is reach out to bernd and lxj to maybe explain what i was hoping to do with that etherpad | 16:00 |
| mrhillsman | it got out of control quickly lol | 16:00 |
| leifmadsen | I say what? lol | 16:00 |
| corvus | mrhillsman: we can discuss the specifics in code review pretty easily, but if we want to schedule a call with folks to talk about the notes in the etherpad, that works too. that might work for getting the big picture. | 16:01 |
| corvus | leifmadsen: important sensible things :) | 16:01 |
| leifmadsen | sounds wrong | 16:01 |
| leifmadsen | I assume you meant someone else | 16:01 |
| mrhillsman | i got to the section about zuul and nodepool and started just taking some quick notes via ## to come back and do the same as i was doing with the environment setup section | 16:02 |
| mrhillsman | yeah | 16:02 |
| mrhillsman | i think getting the big picture is important | 16:02 |
| mrhillsman | because some of those changes you commented on i think were done by bernd because of being specific to his setup | 16:02 |
| pabelanger | Shrews: what size was your all-in-one instance? So far 8GB seems to work in my local testing, with a large HDD | 16:03 |
| pabelanger | 8GB ram I should say | 16:03 |
| Shrews | pabelanger: 64GB :) | 16:03 |
| pabelanger | HA | 16:03 |
| mrhillsman | hehe | 16:03 |
| Shrews | it's my monster machine | 16:03 |
| mrhillsman | i use 3 tiny instances | 16:03 |
| mrhillsman | just like to keep things separate; want to separate them even more | 16:04 |
| pabelanger | ++ | 16:04 |
| pabelanger | ansible makes that easy with group_vars too | 16:05 |
| mrhillsman | corvus i will reply to the review | 16:05 |
| Shrews | i was originally going to do vms (1 per service), but i gave up on that pretty quickly | 16:05 |
| mrhillsman | and i think there is a thread on zuul-discuss so i can respond to that | 16:05 |
| mrhillsman | ah, it's not that bad :) | 16:06 |
| pabelanger | corvus: I'd also love to discus our playbook / role story for zuul / nodepool deployments too. When every you think that is appropriate. | 16:06 |
| mrhillsman | i am using containers right now and they are working fine | 16:06 |
| corvus | i could see us adding a note saying "open these ports if you want to run the servers on separate hosts". but i definitely think aio is a legit use-case, and we don't wan to scare new folks off by giving them the impression they need to run on an openstack scale server farm. :) | 16:06 |
| pabelanger | mrhillsman: which container tech? | 16:07 |
| mrhillsman | plain old docker via lxd | 16:07 |
| mrhillsman | sorry, plain old lxc via lxd | 16:07 |
| mrhillsman | or whatever lol | 16:07 |
| mrhillsman | lxc | 16:07 |
| Shrews | corvus: i'm surprised we'd recommend an aio, but i guess we've never really mapped out how much resources are required. i just assumed "a lot" | 16:07 |
| pabelanger | ah, I've had some early success using this http://git.openstack.org/cgit/openstack/openstack-ansible-plugins/tree/connection/ssh.py plugin that OSA used to manage lxc / systemd-nspawn containers on a remote nodes with ansible. | 16:07 |
| pabelanger | just add a few magic variables into an inventory file, and now things are installed into system containers on the remove node | 16:08 |
| pabelanger | remote* | 16:08 |
| mrhillsman | maybe we can address that with something like | 16:08 |
| mrhillsman | aio setups are perfectly valid even for production | 16:09 |
| corvus | Shrews: not much, actually -- it scales with the size of the system. i think a single 8G vm should be plenty for a 'workgroup-sized' deployment | 16:09 |
| *** acozine1 has joined #zuul | 16:09 | |
| corvus | Shrews: but yeah, we should quantify that :) | 16:09 |
| corvus | maybe i still have my server up... | 16:09 |
| mrhillsman | right | 16:09 |
| Shrews | perhaps we should say that somewhere | 16:09 |
| pabelanger | mrhillsman: yah, I've been testing with single node, but into virtualenvs for services. Not production, but works well in my local 3rd party CI | 16:09 |
| pabelanger | mrhillsman: http://paste.openstack.org/show/717920/ is an example inventory file to create gear01 in a lxc container on bm01 | 16:10 |
| pabelanger | using the connection plugin above | 16:10 |
| mrhillsman | nice | 16:11 |
| pabelanger | OSA has some pretty nice things around lxc | 16:11 |
| corvus | Shrews: when i went through the ZfS guide, i did so on a 4G vm which currently has 1G of ram used. | 16:11 |
| corvus | that's running scheduler, web, executor. | 16:11 |
| Shrews | it's the java things (gerrit, zk) that i wondered about | 16:12 |
| corvus | oh, also running zk | 16:12 |
| corvus | it is not running gerrit | 16:12 |
| pabelanger | what are thought of using review-dev.o.o or gerrithub.io for gerrit example bits? Rather then deployment guide of gerrit? | 16:13 |
| corvus | pabelanger: that could end up with folks stepping on each others toes, since it's not easy to create new projects | 16:14 |
| pabelanger | didn't think of that | 16:15 |
| SpamapS | Shrews: regarding your docs split up.. I'd basically like to have two versions.. 1 "the distro way" and 2 "the container way" | 16:15 |
| SpamapS | I'd really like to move my Zuul into Kubernetes at some point. | 16:16 |
| mrhillsman | ^ | 16:18 |
| mrhillsman | well, not k8s necessarily hehe | 16:18 |
| Shrews | pabelanger: also, there are important pieces in setting up gerrit that we don't document anywhere (ACLs, pipeline label mapping) | 16:18 |
| mrhillsman | but would love to be able to deploy zuul via containers | 16:18 |
| pabelanger | Shrews: traditionally that's been https://docs.openstack.org/infra/system-config/gerrit.html | 16:19 |
| mrhillsman | i think it is ok to remove github from ZfS | 16:19 |
| pabelanger | but agree, we docs for gerrit | 16:20 |
| mrhillsman | and put at top of github page | 16:20 |
| mrhillsman | and then be like, at this point your zuul is ready to do its job | 16:21 |
| Shrews | pabelanger: that really didn't help me much for my setup. we should have something internal to zuul docs | 16:21 |
| mrhillsman | next steps - connector (link link link), reporter (link link link), etc whatever, but your ZfS is done | 16:21 |
| corvus | SpamapS: i have a todo item to write up how we'll build images; once we have that, we can start building docs around it. hopefully now that we released, i'll have time to do that writeup next week. | 16:22 |
| clarkb | Shrews: maybe there is an upstream gerrit doc we can link to? (I've not actually looked) | 16:22 |
| corvus | mrhillsman: well, the idea was to walk all the way through getting a job to run and report, somewhere. | 16:22 |
| Shrews | clarkb: for... ? | 16:22 |
| clarkb | Shrews: for installing gerrit | 16:22 |
| corvus | Shrews, clarkb: last i looked, upstream gerrit's "here's how to set this up in dev" was really good and short. | 16:23 |
| Shrews | clarkb: oh, installing gerrit was not the problem | 16:23 |
| clarkb | Shrews: then zuul only has to deal with make sure stream events acl is set for your user and so on | 16:23 |
| mrhillsman | gotcha | 16:23 |
| corvus | Shrews, clarkb: last i looked was 7 years ago. | 16:23 |
| Shrews | it was linking gerrit and zuul that was the issue | 16:23 |
| corvus | Shrews, clarkb: i assume it's only gotten better and shorter since then :) | 16:23 |
| clarkb | Shrews: gotcha | 16:23 |
| corvus | SpamapS: btw, we released. welcome to the party! | 16:24 |
| * Shrews celebrates with lunch. bbl | 16:24 | |
| *** flepied__ has joined #zuul | 16:25 | |
| *** flepied_ has quit IRC | 16:27 | |
| SpamapS | corvus: I saw, congrats, and thanks, and hooray! | 16:30 |
| *** JasonCL has quit IRC | 16:36 | |
| *** JasonCL has joined #zuul | 16:38 | |
| *** JasonCL has quit IRC | 16:43 | |
| tobiash | corvus, jlk: created a pr to fix that new github3.py issue: https://github.com/sigmavirus24/github3.py/pull/817 | 16:48 |
| *** JasonCL has joined #zuul | 16:49 | |
| *** sshnaidm has quit IRC | 16:50 | |
| corvus | tobiash: ++ | 16:50 |
| *** JasonCL has quit IRC | 16:55 | |
| *** JasonCL has joined #zuul | 16:56 | |
| *** JasonCL has quit IRC | 16:57 | |
| openstackgerrit | Merged openstack-infra/nodepool master: Add host-key-checking option to openstack providers https://review.openstack.org/556624 | 17:00 |
| fdegir | congrats all! | 17:04 |
| pabelanger | fdegir: thanks! | 17:05 |
| fdegir | just saying, I will be coming back with more questions and hopefully contributions as well | 17:07 |
| fdegir | I can't say how much important zuul is | 17:07 |
| corvus | fdegir: thanks! i love both of those! | 17:07 |
| *** JasonCL has joined #zuul | 17:08 | |
| *** JasonCL has quit IRC | 17:13 | |
| rcarrillocruz | folks, on PTO this week | 17:15 |
| rcarrillocruz | congrats on v3 release, great work. Ansible network salutes you! | 17:15 |
| rcarrillocruz | mordred: does it mean you got all the zuul-web stuff in? haven't followed latest changes | 17:15 |
| clarkb | rcarrillocruz: I think the api changing stuff all got it | 17:16 |
| clarkb | *got in | 17:16 |
| clarkb | rcarrillocruz: but there are other items that weren't going to rpesent compatibility issues that were deferred | 17:16 |
| rcarrillocruz | ok | 17:16 |
| mordred | rcarrillocruz: what clarkb said | 17:16 |
| *** JasonCL has joined #zuul | 17:18 | |
| *** JasonCL has quit IRC | 17:19 | |
| *** elyezer has quit IRC | 17:19 | |
| pabelanger | rcarrillocruz: host-key-checking for nodepool above ^ | 17:22 |
| *** flepied__ has quit IRC | 17:22 | |
| *** JasonCL has joined #zuul | 17:22 | |
| rcarrillocruz | ah nice | 17:23 |
| *** JasonCL has quit IRC | 17:30 | |
| *** elyezer has joined #zuul | 17:32 | |
| *** JasonCL has joined #zuul | 17:41 | |
| *** JasonCL has quit IRC | 17:47 | |
| *** JasonCL has joined #zuul | 17:56 | |
| *** JasonCL has quit IRC | 17:58 | |
| *** JasonCL has joined #zuul | 17:59 | |
| *** flepied__ has joined #zuul | 18:01 | |
| *** harlowja has joined #zuul | 18:01 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool master: Reduce logging in _cleanupCurrentProviderUploads function https://review.openstack.org/557791 | 18:17 |
| pabelanger | Shrews: tobiash: corvus: ^noticed on nb01.o.o we are filling INFO logs with those statements. I think dropping them to debug might be more appropriate | 18:18 |
| *** JasonCL has quit IRC | 18:21 | |
| *** mugsie has joined #zuul | 18:22 | |
| clarkb | This channel may be interested as well, I've sent an email to the openstack-infra list with a recap of the cross community infra/ci/cd event from last weekend | 18:26 |
| pabelanger | clarkb: As somebody who didn't attend, thank you | 18:32 |
| clarkb | http://lists.openstack.org/pipermail/openstack-infra/2018-March/005887.html should you not already be subscribed | 18:32 |
| *** sshnaidm has joined #zuul | 18:37 | |
| *** JasonCL has joined #zuul | 18:38 | |
| gundalow | Hi all, where can I see the defination of centos-oci, in particular I'm interested in which version of Python are available | 18:41 |
| gundalow | (nodeset) | 18:42 |
| clarkb | gundalow: it will be in the nodepool config for your zuul. I lack context for what that zuul might be though | 18:42 |
| *** gouthamr has joined #zuul | 18:43 | |
| clarkb | gundalow: in the nodepool config it will either be configured to boot a predefined image or have disk image builder elements listed for building an image in nodepool direclty | 18:43 |
| gundalow | clarkb: ah, yes, the nodeset is just a pointer to a something in nodepool, thanks | 18:44 |
| clarkb | that said isn't centos 7 effectively python2.7 only? | 18:44 |
| clarkb | you can use software collections but any time tripleo brings that up everyone says its not practical | 18:44 |
| *** flepied__ has quit IRC | 18:47 | |
| gundalow | ah, good point. I forgot centos 7 doesn't have py3. I'll stick with what I have for the moment and once I get further on maybe look at building my own image with all the tooling I need | 18:48 |
| gundalow | clarkb: Thanks :) | 18:49 |
| pabelanger | Yah, we do fedora-27 currently to pick up python3 in openstack-infra | 18:49 |
| *** JasonCL has quit IRC | 19:00 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/nodepool master: Reduce logging in _cleanupCurrentProviderUploads function https://review.openstack.org/557791 | 19:00 |
| *** JasonCL has joined #zuul | 19:04 | |
| *** JasonCL has quit IRC | 19:04 | |
| *** JasonCL has joined #zuul | 19:05 | |
| *** JasonCL_ has joined #zuul | 19:06 | |
| *** flepied__ has joined #zuul | 19:08 | |
| *** JasonCL has quit IRC | 19:09 | |
| *** JasonCL_ has quit IRC | 19:11 | |
| clarkb | corvus: I'm going to review the changes to add reno now. Will likely approve them. Please yell if you don't want them approved yet | 19:26 |
| clarkb | shrews acked the nodepool change so decided enough +2s on that one and approved it. Now to the zuul change which looks a bit more involved | 19:29 |
| Shrews | clarkb: *nod* | 19:29 |
| pabelanger | tobiash: do you have an example gate pipeline for github? Would like to compare with SF if possible, trying to debug an issue | 19:30 |
| SpamapS | I have one for GHE | 19:30 |
| *** weshay is now known as weshay_brb | 19:30 | |
| *** weshay_brb is now known as weshay | 19:31 | |
| pabelanger | SpamapS: great | 19:31 |
| SpamapS | pabelanger: http://paste.openstack.org/show/717939/ <-- but I may be weird.. I only gate on the 'approved' label | 19:31 |
| SpamapS | http://paste.openstack.org/show/717940/ <-- those keep that sane | 19:32 |
| clarkb | Shrews: were you interested in also reviewing the zuul change? | 19:32 |
| Shrews | clarkb: just did | 19:32 |
| Shrews | lgtm | 19:32 |
| SpamapS | BTW it would be good to have a way to suppress those utility pipelines from the status page. | 19:32 |
| clarkb | Shrews: oh nice gerrit says we +2'd at the same time | 19:32 |
| clarkb | I'm just going to go ahead and approve thato ne now as I think we have good consensu | 19:33 |
| Shrews | clarkb: we are, like, in sync and stuff | 19:33 |
| Shrews | :-P | 19:33 |
| Shrews | i double approved, in case there was any doubts | 19:33 |
| SpamapS | pabelanger: also 0259b1d083a4ecm is the github user that zuul uses, in case you are wondering. | 19:34 |
| pabelanger | SpamapS: thanks, trying to digest the info | 19:35 |
| pabelanger | really helpful | 19:35 |
| SpamapS | Yeah, there's a bit that is very us-specific there | 19:35 |
| openstackgerrit | Merged openstack-infra/nodepool master: Add reno releasenotes support https://review.openstack.org/536861 | 19:36 |
| pabelanger | SpamapS: thanks, line 16 is what we are missing: http://paste.openstack.org/show/717939/ | 19:42 |
| pabelanger | if I understand, that is applied when somebody clicks approve right? | 19:43 |
| *** gouthamr has quit IRC | 19:44 | |
| *** JasonCL has joined #zuul | 19:45 | |
| *** JasonCL has quit IRC | 19:46 | |
| SpamapS | pabelanger: nope, that is when you add a label named approved | 19:47 |
| SpamapS | pabelanger: if you want to rely on *reviews*, use http://paste.openstack.org/show/717940/ 41-43 | 19:48 |
| SpamapS | pabelanger: we use label, so that people can self-approve | 19:48 |
| SpamapS | since github won't let you approve your own change | 19:48 |
| SpamapS | and then we use that pipeline to add the approved label if an admin approves | 19:48 |
| pabelanger | I see | 19:48 |
| *** JasonCL has joined #zuul | 19:49 | |
| pabelanger | and, crazy question, how can I tell is a PR is approved, by looking at github UI? | 19:49 |
| pabelanger | eg: https://github.com/ansible-network/ansible-zuul-jobs/pull/4 | 19:50 |
| *** JasonCL has quit IRC | 19:51 | |
| *** JasonCL has joined #zuul | 19:55 | |
| *** JasonCL has quit IRC | 19:56 | |
| *** dkranz has quit IRC | 19:57 | |
| pabelanger | okay, I see it now | 19:59 |
| pabelanger | big green check, with <user> approved these changes | 19:59 |
| *** hashar is now known as quibble | 19:59 | |
| pabelanger | thanks, I'll have somebody look into scheduler logs and see what happened | 19:59 |
| *** JasonCL has joined #zuul | 20:00 | |
| *** JasonCL has quit IRC | 20:02 | |
| dmsimard | I know we had discussed the deprecation and removal of a role in zuul-jobs but I forget what we had agreed on | 20:03 |
| clarkb | dmsimard: zuul list next steps email has outline of process for stuff like that | 20:04 |
| dmsimard | oh? I saw the mail but haven't read it yet, looking. | 20:04 |
| dmsimard | doesn't go into the details at least as far as zuul-jobs is concerned | 20:05 |
| * dmsimard search his irc logs | 20:05 | |
| corvus | dmsimard: talking about the ara role? maybe just leave it there -- after all, it's still a valid way to use ara, right? | 20:06 |
| openstackgerrit | Merged openstack-infra/zuul master: Use reno for release notes https://review.openstack.org/262597 | 20:07 |
| dmsimard | corvus: new role is backwards compatible with same defaults as emit-ara-html -- you could say it's a rename of the role because it does more than just html and had additional functionality | 20:08 |
| corvus | dmsimard: ack; then i'd suggest sending a note to zuul-announce, wait a few weeks, then remove. how's that sound? | 20:09 |
| dmsimard | corvus: yeah, I planned on at least including a note in the role README and a debug task to warn about the deprecation notice but I don't expect everyone to actually notice | 20:09 |
| corvus | dmsimard: maybe make the subject really clear, like: Deprecating emit-ara-html role in zuul-jobs | 20:10 |
| corvus | dmsimard: that sounds good too | 20:10 |
| dmsimard | fwiw I was already preparing an email to openstack-dev, doing one for zuul-announce is good too. Does everyone have posting rights there ? | 20:10 |
| corvus | dmsimard: it's fully moderated; the intent is for it to be a communication channel for the software and the zuul-jobs repos, so i think it's appropriate. i'll moderate it through | 20:11 |
| dmsimard | ack | 20:11 |
| *** JasonCL has joined #zuul | 20:12 | |
| pabelanger | tobiash: looks like we are seeing http://paste.openstack.org/show/717856/ on zuul.o.o too, just noticed it by chance | 20:14 |
| *** JasonCL has quit IRC | 20:15 | |
| *** JasonCL has joined #zuul | 20:16 | |
| *** flepied__ has quit IRC | 20:21 | |
| *** flepied has joined #zuul | 20:22 | |
| *** JasonCL has quit IRC | 20:24 | |
| *** JasonCL has joined #zuul | 20:27 | |
| openstackgerrit | David Moreau Simard proposed openstack-infra/zuul-jobs master: Deprecate the emit-ara-html role https://review.openstack.org/557824 | 20:32 |
| *** JasonCL has quit IRC | 20:33 | |
| *** JasonCL has joined #zuul | 20:36 | |
| *** JasonCL has quit IRC | 20:38 | |
| *** JasonCL has joined #zuul | 20:42 | |
| SpamapS | For us, and ARA, I've thought about having our runs feed into a live ARA db | 20:43 |
| SpamapS | right now we're doing the emit-ara-html method, but it would be pretty nice to be able to hit a live ARA and see the ongoing jobs | 20:43 |
| *** JasonCL has quit IRC | 20:47 | |
| openstackgerrit | Mike Fedosin proposed openstack-infra/zuul master: Set default speed_limit and speed_time for the merger https://review.openstack.org/557826 | 20:47 |
| *** quibble has quit IRC | 20:56 | |
| dmsimard | SpamapS: yeah there are (as you know) certain constraints/limitations in doing that with the current version of ARA but it'll be easier with 1.0, promise | 21:02 |
| dmsimard | corvus: how does that sound ? https://etherpad.openstack.org/p/GEOCeAuKhU | 21:02 |
| *** flepied has quit IRC | 21:04 | |
| *** JasonCL has joined #zuul | 21:12 | |
| *** JasonCL has quit IRC | 21:21 | |
| *** JasonCL has joined #zuul | 21:22 | |
| SpamapS | dmsimard: do you want/need help with 1.0? | 21:24 |
| dmsimard | SpamapS: wouldn't turn that down at all | 21:25 |
| dmsimard | SpamapS: I'm working on settling the new 1.0 stuff in place, happy to chat about it sometime next week -- planning some time during this 4 day weekend | 21:26 |
| dmsimard | tl;dr, I threw away ~6 months of on/off work I had put towards 1.0 and went with a different technology stack | 21:26 |
| *** JasonCL has quit IRC | 21:30 | |
| *** JasonCL has joined #zuul | 21:36 | |
| *** JasonCL has quit IRC | 21:37 | |
| *** elyezer has quit IRC | 22:01 | |
| *** elyezer has joined #zuul | 22:02 | |
| *** dvn has joined #zuul | 22:07 | |
| *** acozine1 has quit IRC | 22:22 | |
| dmsimard | corvus: I ended up sending a slightly different version to zuul-announce since I'm EOD, feel free to moderate it if there's anything wrong with it | 22:35 |
| openstackgerrit | Merged openstack-infra/zuul-jobs master: Deprecate the emit-ara-html role https://review.openstack.org/557824 | 22:48 |
| corvus | dmsimard: lgtm, thanks! | 22:56 |
| *** ChanServ sets mode: +c | 22:57 | |
| *** ChanServ changes topic to "Discussion of the project gating system Zuul | Website: https://zuul-ci.org/ | Docs: https://zuul-ci.org/docs/ | Source: https://git.zuul-ci.org/ | Channel logs: http://eavesdrop.openstack.org/irclogs/%23zuul/" | 22:58 | |
| *** JasonCL has joined #zuul | 23:06 | |
| *** JasonCL has quit IRC | 23:07 | |
| clarkb | corvus: https://storyboard.openstack.org/#!/story/2001392 can be closed I Think | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
