| *** armstrongs has joined #zuul | 00:02 | |
| *** armstrongs has quit IRC | 00:11 | |
| *** rfolco has quit IRC | 00:43 | |
| *** y2kenny has quit IRC | 00:48 | |
| *** cdearborn has quit IRC | 01:13 | |
| *** weshay|ruck is now known as weshay | 01:31 | |
| *** swest has quit IRC | 01:34 | |
| *** swest has joined #zuul | 01:49 | |
| *** Goneri has quit IRC | 01:50 | |
| *** ysandeep|away is now known as ysandeep|rover | 01:57 | |
| *** zxiiro has quit IRC | 02:15 | |
| *** ysandeep|rover is now known as ysandeep|rover|b | 03:26 | |
| *** bhavikdbavishi has joined #zuul | 03:39 | |
| *** igordc has quit IRC | 03:42 | |
| *** bhavikdbavishi has quit IRC | 03:48 | |
| *** bhavikdbavishi has joined #zuul | 03:56 | |
| *** bhavikdbavishi1 has joined #zuul | 04:11 | |
| *** bhavikdbavishi has quit IRC | 04:13 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 04:13 | |
| *** ysandeep|rover|b is now known as ysandeep|rover | 04:19 | |
| *** evrardjp has quit IRC | 04:36 | |
| *** evrardjp has joined #zuul | 04:37 | |
| *** swest has quit IRC | 04:55 | |
| *** swest has joined #zuul | 04:55 | |
| *** hashar has joined #zuul | 04:58 | |
| *** threestrands has joined #zuul | 05:06 | |
| *** mnasiadka has quit IRC | 06:21 | |
| *** mnasiadka has joined #zuul | 06:22 | |
| *** avass has joined #zuul | 06:50 | |
| *** sugaar has joined #zuul | 06:53 | |
| *** frickler_ is now known as frickler | 07:05 | |
| *** hashar has quit IRC | 07:06 | |
| *** jcapitao has joined #zuul | 07:13 | |
| *** rpittau|afk is now known as rpittau | 07:17 | |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Connect fingergw to Zookeeper https://review.opendev.org/716875 | 07:27 |
|---|---|---|
| *** masterpe has quit IRC | 07:29 | |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Connect fingergw to Zookeeper https://review.opendev.org/716875 | 07:30 |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Render buildset progress bar correctly https://review.opendev.org/716878 | 07:32 |
| openstackgerrit | Jan Kubovy proposed zuul/zuul master: Connect fingergw to Zookeeper https://review.opendev.org/716875 | 07:40 |
| *** bhavikdbavishi has quit IRC | 07:40 | |
| *** ysandeep|rover is now known as ysandeep|rover|l | 07:43 | |
| *** jpena|off is now known as jpena | 07:54 | |
| *** masterpe has joined #zuul | 07:58 | |
| openstackgerrit | Simon Westphahl proposed zuul/zuul master: Render buildset progress bar correctly https://review.opendev.org/716878 | 08:00 |
| *** threestrands has quit IRC | 08:01 | |
| *** tosky has joined #zuul | 08:02 | |
| AJaeger | I now have a memory error on pep8: https://zuul.opendev.org/t/openstack/build/23286cf5ecc54fccb8d87172ecc54768 | 08:17 |
| *** bhavikdbavishi has joined #zuul | 08:17 | |
| AJaeger | The flake8wrap script looked in .tox etc ;( now fixed ^ | 08:28 |
| *** ysandeep|rover|l is now known as ysandeep|rover | 08:33 | |
| *** hashar has joined #zuul | 08:35 | |
| openstackgerrit | Antoine Musso proposed zuul/zuul master: Enhance some logging messages https://review.opendev.org/704331 | 08:45 |
| openstackgerrit | Antoine Musso proposed zuul/zuul master: Add details to waitUntilSettled() debug log https://review.opendev.org/704380 | 08:48 |
| openstackgerrit | Antoine Musso proposed zuul/zuul master: tests: throttle after all builds have reported https://review.opendev.org/704381 | 08:49 |
| openstackgerrit | Antoine Musso proposed zuul/zuul master: tox: rename pep8 to linters https://review.opendev.org/703635 | 08:58 |
| openstackgerrit | Antoine Musso proposed zuul/zuul master: tox: do not install bindep for linters https://review.opendev.org/703636 | 08:59 |
| *** sshnaidm|afk is now known as sshnaidm | 09:25 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Ensure correct cleanup on repo update and reset https://review.opendev.org/701531 | 10:05 |
| *** rpittau is now known as rpittau|bbl | 10:14 | |
| *** hashar is now known as hasharLunch | 10:17 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Ensure correct cleanup on repo update and reset https://review.opendev.org/701531 | 10:18 |
| *** gtema has joined #zuul | 10:21 | |
| gtema | I'm trying to finally enable gating for github, where I want zuul to merge PR from feature branch into master. Current failure I get is "No base commit found ...". What can I do? | 10:23 |
| openstackgerrit | Merged zuul/zuul-jobs master: Do not end host if correct go version is installed https://review.opendev.org/716607 | 10:24 |
| openstackgerrit | Merged zuul/zuul-jobs master: go-jobs: improve testing https://review.opendev.org/716668 | 10:24 |
| *** ysandeep|rover is now known as ysandeep|brb | 10:34 | |
| *** rfolco has joined #zuul | 11:01 | |
| *** jcapitao is now known as jcapitao_lunch | 11:03 | |
| *** bhavikdbavishi has quit IRC | 11:13 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul master: Ensure correct cleanup on repo update and reset https://review.opendev.org/701531 | 11:19 |
| tobiash | gtema: that sounds lie the feature branch is disconnected from the master | 11:22 |
| tobiash | s/lie/like | 11:22 |
| gtema | it's not, since the github itself can merge it without problems | 11:22 |
| gtema | and this whole branch is like 'git checkout -b FEATURE && git add .zuul.yaml && git push origin FEATURE | 11:23 |
| gtema | so really a single commit | 11:23 |
| tobiash | gtema: in this case a log would help to see where in the workflow this happens | 11:23 |
| gtema | the scheduler fails with message about cross-repo conflict (which it is not), but I see this is anyway a default error | 11:23 |
| gtema | log: this is exactly weird - there are lot's of those "can not find base" debug messages, then it still checks out everything in the tenant, starts some playbooks and there is no real "failure" in the executor log (no separate merger) | 11:24 |
| *** ysandeep|brb is now known as ysandeep|rover | 11:25 | |
| gtema | I mean a really lot output | 11:25 |
| gtema | those merger messages were already hard to find | 11:25 |
| gtema | and no ansible failed plays | 11:26 |
| gtema | I still have a feeling that it tries to do something with other repo (as a cross-dep, which should not be the case). Otherwise I can't understand why it tries to start jobs from totally independent projects | 11:27 |
| *** jpena is now known as jpena|lunch | 11:36 | |
| gtema | tobiash - found it. Very epic and weird: on github master branch policy required linear history (without merge commits) | 12:08 |
| gtema | 1) before scheduler restart I didn't see that in logs, that it was forbidden | 12:09 |
| gtema | 2) github itself is still allows admin to merge it, so I was never actually noticing that | 12:09 |
| *** hasharLunch is now known as hashar | 12:11 | |
| *** bhavikdbavishi has joined #zuul | 12:12 | |
| *** rpittau|bbl is now known as rpittau | 12:12 | |
| *** rlandy has joined #zuul | 12:20 | |
| *** bhavikdbavishi1 has joined #zuul | 12:25 | |
| *** bhavikdbavishi has quit IRC | 12:26 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 12:26 | |
| *** jcapitao_lunch is now known as jcapitao | 12:27 | |
| *** sgw has quit IRC | 12:35 | |
| *** jpena|lunch is now known as jpena | 12:43 | |
| fungi | that sounds like a pitfall we might want to caution about in the github connection driver docs | 12:44 |
| mordred | gtema: it's possible you may want to configure https://zuul-ci.org/docs/zuul/reference/project_def.html?highlight=merge%20mode#attr-project.merge-mode in this case? I also wonder if there is additional support we could do in the github driver in this case (I don't know what's available in the API) | 12:48 |
| mordred | that failure certainly doesn't sound very informative at the least | 12:49 |
| *** bhavikdbavishi has quit IRC | 12:53 | |
| gtema | mordred: but for github they all map to "merge", so there is no other thing to try out with merge-mode | 13:03 |
| gtema | it just seems to me that GH bypasses own sec limitations from UI | 13:05 |
| mordred | gtema: of course it does | 13:09 |
| gtema | like everything :D | 13:09 |
| mordred | gtema: incidentally, when openstac moved to git 9 years ago, one of the two reasons we did not move to github was that there was no way to remove normal push/merge ability from the admins of a project | 13:10 |
| gtema | yeah, I wanted gerrit, but it changes the flow so "dramatically", that not every developer is ok with that (don't throw stones into me) | 13:11 |
| gtema | and GH is for everybody a regular way of devel | 13:11 |
| * fungi is glad not to be a part of "everybody" in that case | 13:13 | |
| mordred | gtema: I hear you. the early days of openstack involved much fighting about gh vs not-github ... and the unfortunate part is much of the argument in favor of gh at that time did not take the form of technical arguments and was instead a bunch of "but I would personally prefer using gh" | 13:13 |
| gtema | sure, I mean everybody outside of openstack :D | 13:13 |
| mordred | (which makes for an unpleasant argument, because then it's just personal) | 13:13 |
| avass | gtema: you could try squash-merge except that it doesn't delete branches | 13:13 |
| gtema | avass: actually default one didn't delete branch as well | 13:14 |
| avass | I think that's a setting in github, let me check | 13:15 |
| gtema | yeah, there is an option for that in settings | 13:17 |
| gtema | "automatically delete head branches" | 13:17 |
| *** ianychoi_ has joined #zuul | 13:17 | |
| avass | gteme: yeah found it, that's what I mean | 13:17 |
| avass | I'm still trying to get my friends to let me host a gerrit instance for us instead of using github | 13:20 |
| *** ianychoi has quit IRC | 13:20 | |
| *** ianychoi_ is now known as ianychoi | 13:23 | |
| corvus | avass: did you want to send a message to zuul-discuss about the renames? | 13:36 |
| AJaeger | there are a few more changes for the linting work, anybody else wants to review the stack starting at https://review.opendev.org/716263 for sphinx? | 13:38 |
| corvus | AJaeger: was just doing that... | 13:38 |
| AJaeger | great | 13:39 |
| corvus | AJaeger, tobiash: i'm thinking there's enough different spellings of sphinx in there that we might want to respin that :) | 13:39 |
| corvus | i count at least two, zero of them correct. it's enough that it might be confusing for future maintenance ('grep -i sphinx' returns no results...) | 13:40 |
| zbr | i would like to reopen the subject of filters in zuul roles, like the one we used at https://review.opendev.org/#/c/705623/18/filter_plugins/flatten_nested_dict.py | 13:41 |
| zbr | what kind of workaround can we do to make it work? | 13:42 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Support multiple matchers when parsing tox output https://review.opendev.org/716263 | 13:42 |
| corvus | AJaeger, tobiash ^ fixed :) | 13:42 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Don't silently ignore exceptions when parsing tox output https://review.opendev.org/716766 | 13:43 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: Strip source dir from file comments https://review.opendev.org/716264 | 13:43 |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: DNM: Debug sphinx message https://review.opendev.org/716722 | 13:43 |
| corvus | rebased | 13:43 |
| corvus | zbr: where was the previous discussion? | 13:44 |
| zbr | corvus: i think arxcruz asked about that yesterday? | 13:45 |
| corvus | zbr: i don't think i was around for that. can you point me to it? | 13:45 |
| AJaeger | corvus: thanks! | 13:45 |
| arxcruz | zbr: corvus yeah, i spoke with AJaeger | 13:45 |
| arxcruz | here | 13:45 |
| arxcruz | basically, a test i'm doing with include role is failing because we have a python filter in our role | 13:45 |
| arxcruz | and it's not possible to use filter outside the trusted jobs | 13:46 |
| fungi | i somehow missed arxcruz mentioning it yesterday. we normally discuss each filter we whitelist, and identify any ways it might be able to write to arbitrary file locations or execute arbitrary code so we can shadow those bits | 13:46 |
| zbr | if i understood corectly, we cannot use custom filters in zuul roles due to security issues. | 13:46 |
| corvus | maybe we could find an eavesdrop link? | 13:46 |
| corvus | so i can catch up? | 13:46 |
| zbr | that is a very simple filter, i guess a workaround would be to make it part of zuul? | 13:46 |
| corvus | it's not in anyone's interest to just type what they typed yesterday again | 13:46 |
| corvus | arxcruz: is this it? http://eavesdrop.openstack.org/irclogs/%23zuul/%23zuul.2020-04-01.log.html#t2020-04-01T11:53:14 | 13:47 |
| AJaeger | corvus: it was no real discussion, we can recap. I asked him to come back again ;) | 13:47 |
| AJaeger | corvus: yeah, time looks ok - there was a link to a failure that is relevant | 13:48 |
| arxcruz | AJaeger: corvus yeah, i just wondering why was failing, asked here, and AJaeger point me that i'm using a filter that is blocked by zuul due secure reasons | 13:48 |
| arxcruz | i ping zbr since he was the one who implement the filter in our collect-logs role | 13:49 |
| zbr | basically is a filter that transforms a nested dict into a list, so ansible can loop over it. i ended up writing it after wasting two days trying to do the same in jinja2 | 13:54 |
| corvus | okay, i think i'm caught up. zbr, arxcruz: zuul doesn't allow untrusted projects to run arbitrary code on its executors, and any kind of ansible plugin can run arbitrary code. i'm not sure there's a good workaround here. putting on my opendev hat, we can't make that a trusted repo. i don't think we want to add arbitrary ansible plugins to zuul. | 13:55 |
| *** weshay is now known as weshay|ruck | 13:55 | |
| corvus | you could write an ansible module to do that; it'd run on the remote host, which is probably excessive, but should work | 13:56 |
| zbr | corvus: i am having a bit of a laugh as I know what ansible 2.10 is cooking regarding all those modules moved from core into community.general collection :D | 13:56 |
| zbr | but i guess that zuul will pre-install community.general collection. | 13:57 |
| *** avass has quit IRC | 13:57 | |
| mordred | zbr: I believe zuul will install ACD | 13:57 |
| *** avass has joined #zuul | 13:57 | |
| mordred | which should have the same effect | 13:57 |
| zbr | yep | 13:58 |
| zbr | now, back at our use-case. what can we do? | 13:58 |
| avass | corvus: yeah I'll do it in a moment | 13:58 |
| zbr | mordred: i could run a shell python command that does the transformation, return json, load json with ansible. ugly like s*** but it should work | 14:00 |
| corvus | zbr: or you could write an ansible module to do that; it'd run on the remote host, which is probably excessive, but should work | 14:01 |
| zbr | the irony is that it will not be safer in any way, once you run shell/python on the localhost, you still get same security issues. | 14:01 |
| *** Goneri has joined #zuul | 14:01 | |
| corvus | zbr: it won't run on localhost | 14:01 |
| corvus | zbr: are you able to receive messages from me? | 14:01 |
| corvus | zbr: as i mentioned, twice, it will run on the *remote* host | 14:02 |
| fungi | zuul doesn't allow you to run arbitrary shell or python on executors | 14:02 |
| zbr | aha, i did not know that. | 14:02 |
| fungi | at least not from untrusted repos | 14:02 |
| zbr | so basically this protection breaks any filters, because by design filters run on controller, but does not break modules. | 14:03 |
| zbr | no problem, I know how to make it a module. | 14:03 |
| zbr | in fact it was originally a module: p | 14:03 |
| mordred | zbr: then it should be quite easy to make a module :) | 14:03 |
| zbr | thanks everyone on this. lesson learn: ansible filters can be tricky with zuul. | 14:04 |
| corvus | zbr: yes -- and it's actually any kind of plugin (not just filters) | 14:04 |
| *** jpena is now known as jpena|off | 14:10 | |
| *** jpena|off is now known as jpena | 14:12 | |
| *** y2kenny has joined #zuul | 14:16 | |
| y2kenny | I have a few questions about nodepool. What are the images that the builder build? Do those images apply generally to all driver or just openstack? (I am not fmiliar with dib images.) | 14:17 |
| tristanC | y2kenny: it's currently only building image for openstack | 14:18 |
| tristanC | (building and uploading) | 14:18 |
| avass | corvus: alright, I sent something. Anyway, three of those changes are still getting "depends on a change with invalid configuration" | 14:20 |
| avass | not sure if I'm missing something or if the error could be what you suggested yesterday | 14:21 |
| y2kenny | I see. Thanks tristanC. Second question: Is there any facility to hook into the lifecycle of nodepool like clean up and tear down? For example, nodepool automatically create namespaces in the k8s (and I assume openshift.) is it possible for nodepool to prepopulate the namespace outside of any job or does the pre and post have to be part of | 14:21 |
| y2kenny | specific jobs? | 14:21 |
| fungi | nodepool doesn't really know anything about jobs. the nodepool launcher gets node requests from the zuul scheduler and then fulfils them, and later removes them once zuul releases its lock on them | 14:22 |
| corvus | y2kenny: it has to be part of a job, but you can make a parent job to do that, and have all jobs that you want to use that behavior inherit from that parent job | 14:22 |
| corvus | y2kenny: job inheritance playbook ordering is like an onion, so parent job pre-playbooks run before children, and parent post-playbooks run after children. | 14:24 |
| y2kenny | ok. What I am thinking is more of a time optimization thing. Having bring up and tear down outside of the job that is general to all jobs can make the CI appear to be more responsible even though the actual work done is the same | 14:25 |
| *** sshnaidm is now known as sshnaidm|afk | 14:25 | |
| corvus | y2kenny: understood. that is a drawback of this. the advantage, however, is that we only have to maintain one system for running remote code on nodepool nodes, not two. :) | 14:25 |
| *** ysandeep|rover is now known as ysandeep|away | 14:26 | |
| y2kenny | ok. Third question: are there facility in nodepool to allow for automatic registration of nodes? or does all nodes have to start with a definition at nodepool (be it static or other kind of provider?) | 14:28 |
| y2kenny | (actually... does nodepool have some kind of rest API?.... I didn't realize to check... let me see...) | 14:29 |
| corvus | y2kenny: no; has to be through a provider. | 14:29 |
| corvus | y2kenny: the idea is that nodepool isn't supposed to be a cloud system, but should interface with them. so if there's some other system that handles registration (like ironic for bare metal?) then we should write a provider to interface with it | 14:30 |
| tristanC | y2kenny: it doesn't have a rest API to push configuration, but it does read and load its configuration file change automatically, so you could scp configuration update to push provider configuration automatically | 14:32 |
| AJaeger | tobiash, corvus, 716766 still shows "Job tox-py37: unable to map line for file comments: etc." | 14:34 |
| y2kenny | corvus: ok, so it's not there by design. baremetal is certainly a use case (I haven't figured out what to do with that just yet. So far I am just using Cobbler to bring up baremtal nodes. I have looked into Ironic may be 3~4 years ago. And now there's metal3, which is based on Ironic.) | 14:34 |
| AJaeger | mordred: ^ | 14:34 |
| mnaser | AJaeger: we can probably do testing now on that stuff | 14:39 |
| y2kenny | tristanC: that's good to know, thanks. I am still trying to figure out the performance side of things because having local on disk code cache is something we have done for our existing infrastructure. I might need to do a lot of pre-provisioning of nodes and having something more dynamic to augment the static provider would be useful. | 14:40 |
| tobiash | corvus: thanks for fixing my stupid typos :) | 14:41 |
| corvus | y2kenny: we do a lot of caching too, and that's definitely something we want to support in nodepool. we have the diskimage builder for use with vm clouds (openstack only now, but could be extended for others). we could also use diskimage builder for bare metal images. k8s i'd probably just build container images in zuul. | 14:42 |
| tobiash | corvus: shall we merge those or shall I add test cases first? | 14:42 |
| corvus | tobiash: maybe at this point adding some testing to the end of the stack would be good | 14:45 |
| tobiash | k | 14:45 |
| tobiash | then I'll -2 the stack until tests are there to prevent accidental merge | 14:45 |
| AJaeger | tobiash: any idea where the warnings on 716766 come from? | 14:47 |
| tobiash | AJaeger: weird, that should have been filtered by the fix yesterday | 14:49 |
| tobiash | AJaeger: oh, it's matching a real 'line comment': /usr/local/lib/python3.6/dist-packages/tox/config/__init__.py:595: UserWarning: conflicting basepython version (set 3.6, should be 3.7) for env 'py37';resolve conflict or set ignore_basepython_conflict | 14:50 |
| tobiash | I think I we should ignore all absolute paths that are left after https://review.opendev.org/716264 | 14:51 |
| AJaeger | good idea | 14:51 |
| y2kenny | corvus: is there something more 'light weight' than diskimage I wonder. I don't really have a solution just yet but what I was thinking initially is to have some kind of standard path on the physical nodes to store ccache or git bare repos. I don't necessarily need the cache to be synced across all nodes. Just something being there to speed | 14:53 |
| y2kenny | things up opportunistically. | 14:53 |
| tristanC | y2kenny: couldn't you use a periodic zuul jobs to refresh your cache on the node directly? | 14:54 |
| corvus | y2kenny: regarding somethign more lightweight than dib: what's your target platform? openstack or something else? regarding the cache -- yes, that's exactly how we handle it in opendev, and there are roles in zuul-jobs that can utilize cached repos.. i'll get some links | 14:55 |
| y2kenny | tristanC: I can definitely do that. I think what I am missing is something that persist across jobs. I think AFS or NFS is what was suggested in the past. | 14:56 |
| corvus | y2kenny: here's our diskimage element that we use to cache git repos on our openstack cloud vm images: https://opendev.org/openstack/project-config/src/branch/master/nodepool/elements/openstack-repos | 14:56 |
| y2kenny | but if I have AFS then I don't really need to do periodic sync. I guess I am looking for disk local cache without a network file system. | 14:57 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Ignore absolute paths after stripping work dir https://review.opendev.org/717042 | 14:58 |
| corvus | y2kenny: then https://zuul-ci.org/docs/zuul-jobs/general-roles.html#role-prepare-workspace-git uses the cache in /opt/git | 14:58 |
| y2kenny | corvus: thanks. I will take a look. Right now I am building on top of k8s (that's why I have been asking a lot of question about the namespace, etc.) | 14:58 |
| corvus | y2kenny: if you're using k8s, then i would look at building container images (inside of zuul) with git repos cached in /opt/git | 14:58 |
| corvus | then have nodepool use those container images | 14:59 |
| y2kenny | corvus: I think that's a possible solution. But I will probably have to experiment a bit (first thing that pops into my mind is the possibility of blowing up /var/lib/docker but perhaps the union fs would be able to deal with the incremental addition to the cache quite well.) | 15:03 |
| *** dmellado has quit IRC | 15:14 | |
| *** sshnaidm|afk is now known as sshnaidm | 15:15 | |
| *** ysandeep|away is now known as ysandeep | 15:18 | |
| avass | corvus, mordred: looks like there are callbacks that can only be configured by environment variables: https://docs.ansible.com/ansible/latest/plugins/callback/jabber.html | 15:20 |
| mordred | avass: "neat" | 15:22 |
| avass | :) | 15:23 |
| *** gtema has quit IRC | 15:28 | |
| *** dmellado has joined #zuul | 15:30 | |
| corvus | avass: wow. well, if we need/want to support those, we should be able to add env variables to zuul.conf and set them when calling ansible | 15:37 |
| avass | corvus: I would guess that there could be custom callbacks that for some reason support both | 15:39 |
| avass | s/support/require | 15:39 |
| *** bhavikdbavishi has joined #zuul | 15:40 | |
| fungi | or propose fixes upstream for those callbacks | 15:40 |
| avass | yeah | 15:40 |
| *** bhavikdbavishi1 has joined #zuul | 15:43 | |
| *** bhavikdbavishi has quit IRC | 15:44 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 15:44 | |
| avass | here's a quick draft what it could look like if we don't want to support environment variables: http://paste.openstack.org/show/791546/ | 15:46 |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Add support for RedHat platforms on install-podman https://review.opendev.org/716578 | 15:48 |
| tristanC | avass: lgtm, isn't callback_name equivalent to the [ansible_callback "name"] ? | 15:49 |
| avass | tristanC: not sure, but what I've seen so far yes | 15:49 |
| corvus | avass: ++ | 15:50 |
| *** smcginnis has left #zuul | 15:53 | |
| *** jamesmcarthur has joined #zuul | 15:55 | |
| avass | tristanC: oh wait, mixed that up with the callback_config_section. Yeah it should probably do that | 15:55 |
| *** rpittau is now known as rpittau|afk | 15:56 | |
| avass | tristanC: but I'm imitating the connection configuration | 15:56 |
| corvus | to be fair, i'm not sure which of [connection "foo"] and name="foo" do what anymore, so if you don't need both, i would recommend not doing both :) | 15:58 |
| corvus | (we should probably figure that out and clean it up for connections) | 15:59 |
| avass | huh, corvus: yeah I was just looking at that | 15:59 |
| avass | corvus: name=foo isn't documented but it's in the example zuul.conf | 15:59 |
| corvus | maybe name=foo doesn't even really exist | 16:00 |
| avass | since the config section needs to be unique anyway i guess [ansible_callback "foo"] is the way to go | 16:00 |
| corvus | yeah, that sounds preferable | 16:00 |
| corvus | tristanC: can we abandon https://review.opendev.org/707908 ? | 16:07 |
| *** jamesmcarthur has quit IRC | 16:11 | |
| *** jamesmcarthur has joined #zuul | 16:12 | |
| corvus | avass: let me know if my comment on https://review.opendev.org/706248 sounds correct | 16:15 |
| *** jamesmcarthur has quit IRC | 16:16 | |
| avass | corvus: I think the idea was something like defaulting to "revoke_sudo: {{ nodepool.attributes.revoke_sudo }}" | 16:17 |
| *** jamesmcarthur has joined #zuul | 16:18 | |
| avass | so it does the same thing, only setting the variable somewhere else | 16:19 |
| *** jamesmcarthur has quit IRC | 16:21 | |
| avass | but I guess the role could do when: "{{ nodepool.attributes.revoke_sudo }}" instead | 16:26 |
| *** ysandeep is now known as ysandeep|away | 16:31 | |
| *** jamesmcarthur has joined #zuul | 16:34 | |
| *** evrardjp has quit IRC | 16:36 | |
| *** evrardjp has joined #zuul | 16:36 | |
| openstackgerrit | Sorin Sbarnea proposed zuul/zuul-jobs master: Add support for RedHat platforms on install-podman https://review.opendev.org/716578 | 16:54 |
| *** bhavikdbavishi has quit IRC | 16:54 | |
| corvus | avass: yeah, i think if we do the second thing, then it means that a site admin can still force revoke sudo on test jobs | 16:57 |
| corvus | avass: because if we allow jobs to turn it off, someone will while we aren't looking :) | 16:57 |
| *** hashar is now known as hasharAway | 17:28 | |
| *** jamesmcarthur has quit IRC | 17:28 | |
| *** jamesmcarthur has joined #zuul | 17:29 | |
| *** jpena is now known as jpena|off | 17:41 | |
| *** ironfoot has quit IRC | 17:43 | |
| *** ironfoot has joined #zuul | 17:43 | |
| avass | corvus: we can do that, I'm just not sure how the node-attributes work yet, since they're only documented on the openstack driver in nodepool | 17:44 |
| avass | https://zuul-ci.org/docs/nodepool/configuration.html#attr-providers.[openstack].pools.node-attributes | 17:44 |
| avass | and I'm not sure if they end up in the jobs yet | 17:45 |
| avass | according to: https://zuul-ci.org/docs/nodepool/releasenotes.html?highlight=node%20attributes#relnotes-3-4-0-new-features they're stored with the node data in zookeeper. But I don't know what the usecase if for that | 17:46 |
| *** zxiiro has joined #zuul | 17:47 | |
| *** jamesmcarthur has quit IRC | 17:54 | |
| *** jamesmcarthur has joined #zuul | 17:55 | |
| *** jcapitao has quit IRC | 17:59 | |
| tobiash | mnaser: the test case for tox is gonna be a bit trickier because of the existing file check :/ | 18:02 |
| Shrews | avass: https://review.opendev.org/#/q/status:open+project:zuul/nodepool+branch:master+topic:node-attr should fix the doc part. Right now, it's only used for zoning executors. | 18:06 |
| Shrews | re: node-attributes, that is | 18:06 |
| corvus | avass, Shrews: yeah, if it isn't passed through all the way to the job, i think we should be able to add that | 18:16 |
| corvus | at least, i can't think of a reason not to right now | 18:16 |
| avass | Shrews: ah nice | 18:16 |
| *** gtema has joined #zuul | 18:17 | |
| Shrews | If anyone wants to take those two changes over for me, feel free. One needs a merge fix, but they should be ready to go otherwise, IMO | 18:23 |
| *** jamesmcarthur has quit IRC | 18:24 | |
| *** sshnaidm is now known as sshnaidm|afk | 18:24 | |
| *** jamesmcarthur has joined #zuul | 18:25 | |
| corvus | Shrews: that first one lgtm; the merge conflict on the second one looks real easy to fix | 18:25 |
| corvus | we can probably just merge the first, then rebase the second | 18:26 |
| *** jamesmcarthur has quit IRC | 18:26 | |
| *** jamesmcarthur has joined #zuul | 18:26 | |
| corvus | tristanC, tobiash: maybe you'd be interested in looking over https://review.opendev.org/714672 ? | 18:30 |
| tobiash | corvus: that was actually on my list but I forgot about that | 18:33 |
| corvus | i'm going to try to get some gce tests soon too | 18:35 |
| *** gtema has quit IRC | 18:50 | |
| mnaser | tobiash: oh that's interesting, maybe the yaml file can be extended to include something like 'files' | 18:55 |
| mnaser | tobiash: actually, that's way more tricky | 18:55 |
| * mnaser mindset is somewhere else but can try and help later | 18:55 | |
| corvus | Shrews: aws has a nice "moto" library for emulating; gcloud doesn't have that, so i think i'll have to make my own emulator like we did for openstack | 19:02 |
| corvus | i might be able to base it on the test from the azure driver | 19:04 |
| *** hasharAway is now known as hashar | 19:14 | |
| *** jamesmcarthur has quit IRC | 19:15 | |
| *** jamesmcarthur has joined #zuul | 19:16 | |
| *** jamesmcarthur has quit IRC | 19:16 | |
| *** jamesmcarthur has joined #zuul | 19:16 | |
| openstackgerrit | Merged zuul/nodepool master: Support node-attributes in static driver https://review.opendev.org/714672 | 19:21 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Update tests for node-attributes https://review.opendev.org/714738 | 19:40 |
| corvus | Shrews: rebased ^ | 19:40 |
| Shrews | corvus: ++ | 19:40 |
| Shrews | thx | 19:40 |
| *** jamesmcarthur has quit IRC | 20:02 | |
| *** jamesmcarthur has joined #zuul | 20:02 | |
| *** jamesmcarthur has quit IRC | 20:19 | |
| *** saneax has quit IRC | 20:21 | |
| *** jamesmcarthur has joined #zuul | 20:26 | |
| tobiash | mnaser: I think I found a way but will finish it probably tomorrow | 20:36 |
| *** weshay|ruck is now known as weshay | 20:57 | |
| *** jamesmcarthur has quit IRC | 21:03 | |
| *** jamesmcarthur has joined #zuul | 21:03 | |
| *** igordc has joined #zuul | 21:05 | |
| *** jamesmcarthur has quit IRC | 21:08 | |
| *** dpawlik has quit IRC | 21:12 | |
| *** rlandy is now known as rlandy|brb | 21:21 | |
| *** jamesmcarthur has joined #zuul | 21:26 | |
| *** jamesmcarthur has quit IRC | 21:26 | |
| *** jamesmcarthur has joined #zuul | 21:26 | |
| mnaser | tobiash: cool, look forward for that :) | 21:32 |
| *** armstrongs has joined #zuul | 21:37 | |
| *** armstrongs has quit IRC | 21:48 | |
| *** hashar has quit IRC | 21:49 | |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add GCE driver tests https://review.opendev.org/717133 | 22:02 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Fix shutdown ordering https://review.opendev.org/717134 | 22:02 |
| corvus | Shrews: ^ there's my penance :) | 22:02 |
| *** rlandy|brb is now known as rlandy | 22:17 | |
| *** openstackstatus has quit IRC | 22:56 | |
| *** openstack has joined #zuul | 23:00 | |
| *** ChanServ sets mode: +o openstack | 23:00 | |
| *** tosky has quit IRC | 23:00 | |
| *** rlandy has quit IRC | 23:11 | |
| *** zxiiro has quit IRC | 23:24 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!