| tristanC | there is also "/usr/bin/python2: No module named wheel" earlier in the failure, so perhaps https://review.opendev.org/#/c/736402/ is the root cause | 00:01 |
|---|---|---|
| ianw | tristanC: i think the epel requirement is a bit of a red herring. if the role is being asked to install python2 pip packages it needs epel. it's more why the role is trying to install those | 00:05 |
| ianw | it's looking if ansible is running under python2, and if it is, assuming you want the python2 packages | 00:08 |
| ianw | i feel like that's a reasonable assumption. you're running under python2 and *don't* want the python2 packages, because you've installed the python3-pip packages in the dockerfile | 00:09 |
| ianw | i'm not sure there's a more sensible default ensure-pip could make that wouldn't break you | 00:10 |
| *** wuchunyang has joined #zuul | 00:22 | |
| tristanC | ianw: that works for me, i guess we just need to document that centos-7 needs epel installed in order to use zuul-jobs with python2. otherwise we'll configure centos-7 node to use python3 ansible interpreter to avoid the new failure. | 00:22 |
| ianw | i agree we can make this clearer, i'm just working on something | 00:23 |
| *** wuchunyang has quit IRC | 00:26 | |
| *** yolanda has quit IRC | 00:47 | |
| *** Goneri has quit IRC | 00:51 | |
| *** yolanda has joined #zuul | 00:51 | |
| *** piotrowskim has quit IRC | 00:59 | |
| openstackgerrit | Ian Wienand proposed zuul/zuul-jobs master: ensure-pip: Clarify situation with ensure_pip_from_packages_with_python2 https://review.opendev.org/742571 | 01:01 |
| ianw | tristanC: ^ I think that clarifies the reality of what's going on | 01:01 |
| *** rlandy|bbl is now known as rlandy | 01:39 | |
| *** _erlon_ has quit IRC | 01:41 | |
| *** decimuscorvinus has quit IRC | 01:46 | |
| *** decimuscorvinus has joined #zuul | 02:12 | |
| *** rfolco has quit IRC | 02:13 | |
| *** decimuscorvinus has quit IRC | 02:16 | |
| *** decimuscorvinus has joined #zuul | 02:16 | |
| *** rfolco has joined #zuul | 02:31 | |
| *** sgw has quit IRC | 02:32 | |
| *** rfolco has quit IRC | 02:40 | |
| *** bhavikdbavishi has joined #zuul | 02:43 | |
| *** sgw has joined #zuul | 02:48 | |
| *** bhavikdbavishi1 has joined #zuul | 02:48 | |
| *** bhavikdbavishi has quit IRC | 02:49 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 02:49 | |
| *** sgw has quit IRC | 02:52 | |
| *** sgw has joined #zuul | 02:57 | |
| openstackgerrit | Ian Wienand proposed zuul/nodepool master: container tests: map devstack TLS CA into containers https://review.opendev.org/742579 | 03:06 |
| *** rlandy has quit IRC | 03:10 | |
| *** bhavikdbavishi has quit IRC | 03:16 | |
| *** bhavikdbavishi has joined #zuul | 03:17 | |
| *** sgw has quit IRC | 03:40 | |
| *** sgw has joined #zuul | 03:40 | |
| *** sgw has quit IRC | 04:15 | |
| *** sgw has joined #zuul | 04:15 | |
| *** bhavikdbavishi has quit IRC | 04:21 | |
| *** bhavikdbavishi has joined #zuul | 04:23 | |
| *** sgw1 has quit IRC | 04:59 | |
| *** sanjayu_ has joined #zuul | 05:20 | |
| *** marios has joined #zuul | 05:22 | |
| *** sanjayu__ has joined #zuul | 05:22 | |
| *** sanjayu_ has quit IRC | 05:25 | |
| *** vishalmanchanda has joined #zuul | 05:39 | |
| *** reiterative has quit IRC | 05:44 | |
| *** reiterative has joined #zuul | 05:44 | |
| *** bhavikdbavishi1 has joined #zuul | 05:55 | |
| *** bhavikdbavishi has quit IRC | 05:57 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 05:57 | |
| *** bhavikdbavishi has quit IRC | 06:46 | |
| *** bhavikdbavishi has joined #zuul | 07:08 | |
| *** nils has joined #zuul | 07:18 | |
| *** wuchunyang has joined #zuul | 07:20 | |
| *** tosky has joined #zuul | 07:22 | |
| *** wuchunyang has quit IRC | 07:29 | |
| *** jcapitao has joined #zuul | 07:43 | |
| *** sugaar has joined #zuul | 07:50 | |
| *** jpena|off is now known as jpena | 07:55 | |
| *** wuchunyang has joined #zuul | 07:59 | |
| *** wuchunyang has quit IRC | 08:01 | |
| *** harrymichal has joined #zuul | 08:48 | |
| *** vishalmanchanda has quit IRC | 09:10 | |
| *** holser has quit IRC | 09:14 | |
| *** holser has joined #zuul | 09:16 | |
| *** holser has quit IRC | 09:16 | |
| *** jpena is now known as jpena|afk | 09:37 | |
| *** hamalq has quit IRC | 09:41 | |
| *** sanjayu__ has quit IRC | 09:51 | |
| *** saneax has joined #zuul | 09:54 | |
| *** bhavikdbavishi has quit IRC | 09:59 | |
| *** bhavikdbavishi has joined #zuul | 10:17 | |
| *** jpena|afk is now known as jpena | 10:22 | |
| *** zbr is now known as zbr|ruck | 10:30 | |
| *** jcapitao is now known as jcapitao_lunch | 10:58 | |
| webknjaz | FYI `cheroot==8.4.0` is out | 10:59 |
| *** harrymichal has quit IRC | 11:07 | |
| *** vishalmanchanda has joined #zuul | 11:09 | |
| *** sshnaidm|afk is now known as sshnaidm|off | 11:11 | |
| *** bhavikdbavishi has quit IRC | 11:35 | |
| *** jpena is now known as jpena|lunch | 11:36 | |
| *** bhagyashris is now known as bhagyashris|afk | 11:54 | |
| *** harrymichal has joined #zuul | 12:05 | |
| *** holser has joined #zuul | 12:12 | |
| *** rlandy has joined #zuul | 12:13 | |
| *** rfolco has joined #zuul | 12:13 | |
| *** jcapitao_lunch is now known as jcapitao | 12:19 | |
| *** bhagyashris|afk is now known as bhagyashris | 12:22 | |
| *** Goneri has joined #zuul | 12:25 | |
| *** jpena|lunch is now known as jpena | 12:39 | |
| *** harrymichal has quit IRC | 12:40 | |
| mnaser | s/o to patternfly 3 work, the new zuul ui looks pretty. | 13:11 |
| *** _erlon_ has joined #zuul | 13:15 | |
| *** Goneri has quit IRC | 13:26 | |
| tristanC | ianw: thank you, that looks great! | 13:27 |
| *** harrymichal has joined #zuul | 13:48 | |
| *** harrymichal has quit IRC | 14:46 | |
| *** harrymichal has joined #zuul | 14:46 | |
| *** sgw1 has joined #zuul | 14:56 | |
| *** bhavikdbavishi has joined #zuul | 15:08 | |
| *** bhavikdbavishi1 has joined #zuul | 15:11 | |
| *** bhavikdbavishi has quit IRC | 15:12 | |
| *** bhavikdbavishi1 is now known as bhavikdbavishi | 15:12 | |
| corvus | okay, i'm going to re-enqueue the release tag | 15:23 |
| corvus | (and yay, it does have the docker build in it this time) | 15:24 |
| corvus | 2020-07-23 15:32:53.616088 | ubuntu-bionic | ModuleNotFoundError: No module named 'setuptools' | 15:34 |
| corvus | https://744f27414b9103148995-ebb0bf58265a2173bda790006658dd60.ssl.cf5.rackcdn.com/756982b0762d8739191e486257ff983aac5448e1/release/zuul-release-python/9e7f55c/job-output.txt | 15:34 |
| corvus | that's the python release failing | 15:34 |
| corvus | anyone grok that? | 15:35 |
| clarkb | I think that is the ensure-pip related stuff. Does Zuul not run a tarball build job? I would expectit and the release to overlap on the build side | 15:37 |
| corvus | it runs "zuul-build-python-release" on check | 15:38 |
| corvus | and that ran okay on the most recent 3.x patch | 15:39 |
| clarkb | ah the failing bit is in opemdev base-jobs | 15:40 |
| clarkb | which is probably why they differ. I bet non release doesnt parent to that | 15:40 |
| corvus | wow yeah that is a different playbook | 15:40 |
| corvus | clarkb: so you suspect one of them has some kind of ensure-pip thing in pre? | 15:42 |
| clarkb | yes | 15:42 |
| corvus | yeah, looks like ensure-pip is run in the check job | 15:43 |
| corvus | clarkb: you think ensure-pip is the only thing missing? | 15:43 |
| corvus | remote: https://review.opendev.org/742694 Add ensure-pip to python release pre-run playbook | 15:44 |
| clarkb | corvus: I think that will address missing setuptools. Looking at the history of build-release-python is an addition of wheel then removal of it | 15:45 |
| clarkb | its not yet clear to mw why it was added and how it broke leading it to be removed | 15:46 |
| clarkb | maybe mordred and zbr remember | 15:46 |
| *** marios is now known as marios|out | 15:55 | |
| *** zbr|ruck is now known as zbr | 16:02 | |
| *** holser_ has joined #zuul | 16:07 | |
| openstackgerrit | James E. Blair proposed zuul/zuul-jobs master: DNM: test lint on ensure-pip https://review.opendev.org/742699 | 16:08 |
| *** holser has quit IRC | 16:09 | |
| *** holser has joined #zuul | 16:11 | |
| *** harrymichal has quit IRC | 16:12 | |
| *** harrymichal has joined #zuul | 16:12 | |
| *** hamalq has joined #zuul | 16:12 | |
| tobiash | that's weird, the task it's complaining about is a shell task where env vars work, not a command task | 16:13 |
| *** holser_ has quit IRC | 16:14 | |
| *** hamalq_ has joined #zuul | 16:15 | |
| *** holser has quit IRC | 16:16 | |
| fungi | yep, discussion has been happening in #opendev | 16:16 |
| *** holser_ has joined #zuul | 16:17 | |
| fungi | since we ran into it on a build for opendev/base-jobs | 16:17 |
| *** hamalq has quit IRC | 16:19 | |
| *** Goneri has joined #zuul | 16:20 | |
| corvus | this is the longest emergency release process ever. :( | 16:23 |
| *** jcapitao has quit IRC | 16:27 | |
| *** holser_ has quit IRC | 16:27 | |
| *** holser has joined #zuul | 16:28 | |
| avass | tobiash: no not using the s3 role yet. I wrote it because I stumbled upon minio and I was bored | 16:33 |
| *** holser has quit IRC | 16:34 | |
| *** holser has joined #zuul | 16:34 | |
| avass | tobiash: we'll probably start using it whenever I get back from vacation though :) | 16:34 |
| avass | zbr: I see no problems with the log wrapping | 16:35 |
| *** holser has quit IRC | 16:39 | |
| *** holser has joined #zuul | 16:40 | |
| corvus | we're going to do log wrapping after the user prefs, right? | 16:49 |
| zbr | thanks. apparently i need to implement the config storage | 16:49 |
| zbr | i think that log wrapping is the most simplistic change that needs config UI | 16:50 |
| zbr | so it is a good candidate to be the first | 16:50 |
| corvus | i'm going to re-enqueue the tag -- i think that will push new docker images, but i think that's okay; clarkb, fungi: agree? | 16:50 |
| corvus | (i'm less worried about that than i am with, say, re-uploading to pypi) | 16:51 |
| clarkb | corvus: yes I expect that is ok | 16:51 |
| corvus | away we go then | 16:51 |
| openstackgerrit | Albin Vass proposed zuul/zuul-jobs master: Update upload-logs-s3 with access/secret key variable documentation https://review.opendev.org/742710 | 16:52 |
| avass | tobiash: I realized I forgot to add those to the README ^ | 16:52 |
| *** marios|out has quit IRC | 17:01 | |
| *** jpena is now known as jpena|off | 17:01 | |
| clarkb | the python release has completed. it is doing docker images now | 17:07 |
| fungi | corvus: sorry, stepped away for a few but reenqueuing the tag should work, yes. sounds good | 17:11 |
| clarkb | corvus: all jobs have succeeded now | 17:19 |
| clarkb | I think that means we should be good for the release | 17:19 |
| corvus | yep, i'll double check pypi / dockerhub | 17:20 |
| corvus | timestamps on dockerhub look good | 17:21 |
| corvus | pypi looks good | 17:22 |
| corvus | tobiash: https://review.opendev.org/742461 if you have a second | 17:24 |
| corvus | i've approved the change to re-add the in-development release notes | 17:24 |
| corvus | i'll prepare the release announcement now | 17:24 |
| *** bhavikdbavishi has quit IRC | 17:27 | |
| *** bhavikdbavishi has joined #zuul | 17:28 | |
| *** vishalmanchanda has quit IRC | 17:28 | |
| corvus | okay, next i think we should delete the stable/3.x branch, and then work on a nodepool release which bumps kazoo? | 17:29 |
| zbr | corvus: maybe you could lead the way and convert zuul_tz_string to use a single config json in localStorage? | 17:29 |
| *** harrymichal has quit IRC | 17:30 | |
| corvus | zbr: that's a good idea, i'll try to find some time to do that; it may not be immediately | 17:30 |
| *** harrymichal has joined #zuul | 17:30 | |
| zbr | i do have a feeling that at any step I take another requirement appears to do something else that was never done | 17:30 |
| corvus | zbr: well, using redux is certainly done | 17:30 |
| corvus | zbr: and the new stuff you want to do is your idea | 17:30 |
| zbr | well, my idea was to wrap logs, i was asked to add preferences. | 17:31 |
| zbr | i did | 17:31 |
| zbr | also to use localStorage instead of cookies, done too. | 17:31 |
| corvus | pretty sure it was your idea to use localstorage :) | 17:31 |
| zbr | well... kudos to me:D | 17:32 |
| corvus | indeed, it's a good idea | 17:32 |
| clarkb | corvus: for nodepool we need to figure out container images builds. https://review.opendev.org/#/c/741973/ is where I got to on that | 17:32 |
| zbr | only because I know how bad are cookies (privacy being the smaller issue with them) | 17:32 |
| corvus | clarkb: are we completely broken there? should we just turn off arm for the moment? | 17:32 |
| clarkb | corvus: well I got it working just with a bunch of compromises. The most recent is using openstack's constraints listings for crpytography and bcrypt and pynacl as they seem to be the most costly packages | 17:33 |
| clarkb | (the timed out jobs there are due to a different issue that ianw was looking at related to our use of tls-proxy in ensure-devstack, there is a revert proposed if we need it) | 17:33 |
| clarkb | corvus: I think it mostly comes down to how comfortable we are with those compromises. If the answer is not very then ya I think we should disable it for now | 17:33 |
| corvus | clarkb: i'm on the side of not-very | 17:34 |
| zbr | i wonder if we have any practical advantages for going blob-config approach or it would be easier to just have each config item with its own store/load. | 17:34 |
| zbr | also our use case seems quite generic, i wonder if we cannot find an already written react component that can do just that, load and save settings. | 17:34 |
| corvus | zbr: mostly comes down to how the redux store is setup | 17:35 |
| corvus | clarkb: i really don't want to be chasing or tied to openstack versions of deps | 17:35 |
| clarkb | corvus: the good news is that I think we've discovered where the problems are. THe bad news is that ya we don't have great options for addressing them right now. And I agree I think it is better for the arm and x86 packages to not need different versions of deps | 17:35 |
| corvus | clarkb: do you think the intermediate dependency layer idea may be helpful? i haven't had a chance to try an implementation of that | 17:36 |
| clarkb | corvus: yes, I think that would also address the problem. What I don't know is whether or not we can build that intermediate layer within our job timeouts (I don't think we successfully built a bcrypt, pynacl or cryptography package on buildx yet due to hitting the current hour timeout on that job) | 17:37 |
| clarkb | basically if ^ can be done within reasonable time period then yes I think it will address the problem for us | 17:37 |
| clarkb | there time period is job runtime not implementation time | 17:37 |
| corvus | clarkb: well, i was thinking it would be a completely new image, so we don't build it unless a dependency changes | 17:38 |
| clarkb | also when new releases happen upstream we'll need to wait for however long that build happens for jobs to work again (this is the upside to the constraints system is we can bump when it will work) | 17:38 |
| clarkb | corvus: ya I know, I'm just thinking if it hits our 3 hour timeout limit then we're stuck again | 17:38 |
| clarkb | corvus: and when releases happen upstream we'd have to wait $time for things to start working again | 17:38 |
| zbr | corvus: i wonder if we should not better implement load/save at app level, and save the list of persisten propTypes, like https://opendev.org/zuul/zuul/src/branch/master/web/src/App.jsx#L81 | 17:38 |
| corvus | clarkb: we could also go back and come up with a different multi-arch approach; one where we build them separately and then combine the manifests | 17:40 |
| corvus | clarkb: all this was predicated on the assumption that the arm build was "a minute or two" slower, which i think was determined under flawed testing | 17:41 |
| clarkb | ya I think cpu intensive activity like compilation in particular is a problem whereas file copies (general io) isn't | 17:41 |
| corvus | clarkb, fungi: i'm about to delete the stable/3.x branch; but technically we'll lose the reference to the last commit on that branch (since it was one past the tag and updated the release job defn). i don't really care and am fine with that; does that sound okay to you too? | 17:42 |
| clarkb | corvus: hrm, if we decide we need a 3.19.2 we'll need to remember to add that back in? | 17:42 |
| corvus | clarkb: i'm sort of inclined to disable the arm builds for now so nodepool isn't stuck, then work on re-enabling them with one of these solutions | 17:43 |
| corvus | clarkb: yep | 17:43 |
| clarkb | re branch cleanup thats my only concern with losing that commit. the commit itself isn't important. If you think we'll remember or that 3.19.2 is very unlikely I think it is ok | 17:43 |
| corvus | and remember to do a bunch of other stuff i don't care to deal with either, like make sure we don't publish docs or latest docker images from the stable branch... it's all messy, and i have no intention of doing another 3.x. :) | 17:43 |
| clarkb | re nodepool images I'm ok with disabling them. we can also enable them speculatively | 17:43 |
| clarkb | probably the most important thing those arm images has done for us is teach us a lot about doing multi arch docker builds | 17:44 |
| clarkb | we now know that the infrastructure can do it, if slowly and where the slow bits are | 17:44 |
| clarkb | additionally I think this experience has shown that it is unlikely that a qemu'd nodepool builder will be performant enough for dib image builds (but we should probably test that if we still think that is a potential option) | 17:47 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Stop building arm container images https://review.opendev.org/742717 | 17:47 |
| corvus | clarkb: that's a good point | 17:47 |
| corvus | looks like the kazoo bump is merged in nodepool | 17:48 |
| corvus | so we just want to merge the disable arm change so that our image builds might actually succeed, then i think we can tag | 17:48 |
| clarkb | corvus: if we need to land other nodepool changes we'll need to address the tls-proxy thing. I'm looking to see if ianw got a proper fix up for that or if we should just go with the revert for now. Those integration jobs don't run on tags so if we just tag we'll be fine | 17:49 |
| clarkb | corvus: https://review.opendev.org/#/c/742579/1 is the proper fix. I think that will succeed when the arm64 images are removed | 17:50 |
| clarkb | so we can land that one if necessary | 17:50 |
| clarkb | (I mean we should land it anyway, but we know where to look if this is a problem) | 17:50 |
| corvus | clarkb: we merged a change we depends-on tested even though the depends-on test failed? :( | 17:52 |
| clarkb | corvus: it failed due to the image builds which masked the other problem :/ | 17:53 |
| corvus | clarkb: proper fix lgtm -- i'm assuming we'll be able to see it work after arm is disabled? | 17:53 |
| clarkb | basically we had two failures but assumed just the known one was present (which was a bad assumption) | 17:54 |
| clarkb | and ya once arm is gone the fix should be good | 17:54 |
| corvus | clarkb: how about i restack it behind the arm disable and we can get that going? | 17:54 |
| clarkb | ++ | 17:54 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: container tests: map devstack TLS CA into containers https://review.opendev.org/742579 | 17:54 |
| corvus | clarkb: is the arm change going to fail on that? | 17:55 |
| clarkb | oh ya :/ | 17:55 |
| corvus | i'll squash then | 17:55 |
| clarkb | k | 17:56 |
| *** bhavikdbavishi has quit IRC | 17:56 | |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Stop building arm container images and fix devstack tls https://review.opendev.org/742579 | 17:56 |
| *** bhavikdbavishi has joined #zuul | 17:56 | |
| fungi | corvus: i'm not concerned about losing the reference to the tip of stable/3.x, it's present on master anyway | 17:58 |
| fungi | (the code change itself, i mean) | 17:59 |
| fungi | also it may not actually disappear? gerrit does have a named ref for the commit at least, though not any merge commit which stitched it in | 18:00 |
| fungi | if there was one | 18:00 |
| corvus | true | 18:00 |
| fungi | in fact, there is no merge commit there | 18:01 |
| *** bhavikdbavishi has quit IRC | 18:01 | |
| corvus | okay, i'm deleting the stable/3.x branch, currently at revision 0530260ca8e1c9719ac503e2d987be276cc1d2e0 | 18:01 |
| fungi | so the named ref for 742541 is also the current tip of stable/3.x | 18:01 |
| corvus | that was fun let's never do it again | 18:01 |
| *** harrymichal has quit IRC | 18:19 | |
| *** harrymichal_ has joined #zuul | 18:19 | |
| *** holser has quit IRC | 18:34 | |
| *** holser has joined #zuul | 18:34 | |
| *** holser has quit IRC | 18:39 | |
| *** holser has joined #zuul | 18:45 | |
| tobiash | corvus: 742461 lgtm | 18:48 |
| tobiash | corvus: maybe next time it's easier to patch reno so it works with multiple branches ;) | 18:55 |
| *** hamalq_ has quit IRC | 18:56 | |
| *** holser has quit IRC | 18:56 | |
| *** holser has joined #zuul | 18:57 | |
| openstackgerrit | Merged zuul/zuul-jobs master: Reinstate docker tag cleanup https://review.opendev.org/742461 | 18:58 |
| tobiash | avass: regarding the s3 upload, was it intentional that it doesn't create the containers if they're not existing? the swift version does (and I think in a multi-tenant environment with container per tenant this would be advantageous) | 18:58 |
| *** holser_ has joined #zuul | 19:05 | |
| *** holser has quit IRC | 19:05 | |
| avass | tobiash: I believe the gcs role doesn't and I just copied the behaviour of that | 19:10 |
| tobiash | ah ok | 19:11 |
| avass | so we could add | 19:11 |
| avass | that behaviour :) | 19:12 |
| tobiash | :) | 19:12 |
| *** holser_ has quit IRC | 19:14 | |
| *** holser has joined #zuul | 19:14 | |
| corvus | yeah, i think there was a whole bunch more stuff that gcs needed to know to create a bucket | 19:16 |
| tobiash | ah that explains it | 19:17 |
| corvus | storage location, access controls, storage policy, lifecycle... | 19:17 |
| corvus | we could probably include all that as input, but it'll take a bit of work | 19:17 |
| *** holser_ has joined #zuul | 19:17 | |
| corvus | really good thing to think about as we think about combining the roles though :) | 19:17 |
| tobiash | out of curiousity, is it possible to spread a module over multiple files (within the same role)? | 19:18 |
| avass | oh, good question actually | 19:19 |
| tobiash | if not, we'd probably need to merge those three modules into one that can do all three based on config | 19:19 |
| *** holser has quit IRC | 19:19 | |
| avass | I would guess that ansible copies the entire library/* to the remote | 19:20 |
| *** holser_ has quit IRC | 19:21 | |
| *** holser has joined #zuul | 19:22 | |
| *** holser_ has joined #zuul | 19:23 | |
| *** holser__ has joined #zuul | 19:25 | |
| tobiash | seems to be doable by putting shared code into ./module_utilrs/foo.py | 19:27 |
| tobiash | https://docs.ansible.com/ansible/latest/dev_guide/developing_module_utilities.html#using-and-developing-module-utilities | 19:27 |
| *** holser has quit IRC | 19:27 | |
| *** holser_ has quit IRC | 19:28 | |
| avass | tobiash: oh, nice | 19:28 |
| *** holser__ has quit IRC | 19:30 | |
| tobiash | so in a first step we could move the custom modules into an upload-logs-base role and add refactor the common parts into module_utils | 19:30 |
| *** holser has joined #zuul | 19:30 | |
| tobiash | the real roles could then mention this role as dependency so they can be unchanged at first | 19:30 |
| *** holser has quit IRC | 19:31 | |
| *** holser has joined #zuul | 19:32 | |
| *** holser_ has joined #zuul | 19:35 | |
| *** holser has quit IRC | 19:36 | |
| avass | sounds like a plan | 19:37 |
| *** holser_ has quit IRC | 19:40 | |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Merge upload logs modules into common role https://review.opendev.org/742732 | 19:46 |
| corvus | we use module_utils in a few places in zuul-jobs | 19:46 |
| corvus | also wow nodepool change still not merged | 19:47 |
| tobiash | yeah the buildset registry uses it | 19:47 |
| openstackgerrit | Merged zuul/nodepool master: Stop building arm container images and fix devstack tls https://review.opendev.org/742579 | 19:49 |
| *** harrymichal_ has quit IRC | 19:50 | |
| *** harrymichal has joined #zuul | 19:50 | |
| clarkb | corvus: ^ I think you should be good to tag now? | 19:52 |
| corvus | oh you know what we forgot? release note | 19:57 |
| corvus | i'll copy the zk note from zuul real quick | 19:57 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Add release note about kazoo upgrade https://review.opendev.org/742735 | 19:59 |
| clarkb | +2 | 19:59 |
| corvus | i +3d it | 20:00 |
| corvus | i guess i'll check back in in a few hours | 20:00 |
| fungi | oh, it should be ready for a recheck before that long ;) | 20:00 |
| openstackgerrit | Merged zuul/zuul-jobs master: Disable base role testing that runs code on localhost https://review.opendev.org/742537 | 20:04 |
| openstackgerrit | Merged zuul/zuul-jobs master: Collect partial subunit files https://review.opendev.org/742527 | 20:04 |
| corvus | clarkb: do we need to do remediation for 742537? | 20:04 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Consolidate common log upload code into module_utils https://review.opendev.org/742736 | 20:06 |
| *** saneax has quit IRC | 20:06 | |
| clarkb | corvus: possibly? some of what was disabled likely has minimal value (in fetch-subunit-role we were asserting html content on localhost) | 20:07 |
| clarkb | corvus: other parts would be nice to test if we can, but I'm not sure how to do that. In particular the workspace setup role | 20:07 |
| clarkb | since that assumes it is running trusted and doing stuff on the execution as well as the remote | 20:07 |
| clarkb | it may be the case that using base-test more often is the reality we operate under? | 20:08 |
| clarkb | I commented things out rather than deleting to make it clearer where the gaps are if we want to try and fix it | 20:08 |
| corvus | well, let's see. i don't care about zuul-cloner. :) | 20:09 |
| fungi | that was also my sentiment regarding z-c | 20:09 |
| fungi | and "please stop using it" | 20:09 |
| corvus | i think that leaves subunit and prepare-workspace git | 20:09 |
| corvus | i think those are in reverse order of importance | 20:09 |
| corvus | so i'll take a look at what's needed to run prepare-workspace-git | 20:10 |
| clarkb | ya and the subunit one still has good coverage | 20:10 |
| clarkb | we basically check that all the normal operations succeed but in doing extra verification of that were running localhost code | 20:10 |
| corvus | (not entirely sure what fetch-subunit-output is doing in base-roles, but maybe there's a reason) | 20:10 |
| corvus | i think the only way to test that would be to use a nested ansible | 20:17 |
| corvus | it's the "git push" from the executor that hits is | 20:17 |
| corvus | it's too bad the ansible git module doesn't support push | 20:17 |
| corvus | maybe we just leave this alone for now, and work on removing the restriction in v4 | 20:18 |
| clarkb | we've also got base-test if we need it | 20:19 |
| clarkb | that role has been pretty stable recently so maybe its a non issue :) | 20:19 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Merge upload logs modules into common role https://review.opendev.org/742732 | 20:25 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Consolidate common log upload code into module_utils https://review.opendev.org/742736 | 20:25 |
| openstackgerrit | Merged zuul/nodepool master: Add release note about kazoo upgrade https://review.opendev.org/742735 | 20:39 |
| clarkb | corvus: ^ only took 40 minutes :) | 20:40 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Merge upload logs modules into common role https://review.opendev.org/742732 | 20:44 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Consolidate common log upload code into module_utils https://review.opendev.org/742736 | 20:44 |
| corvus | clarkb: that is faster than i can write an email | 20:50 |
| corvus | commit c216149325e76e86e9881e2dc195d501af956435 (HEAD -> master, tag: 3.13.1, origin/master, origin/HEAD, refs/changes/35/742735/1) | 20:51 |
| corvus | zuul-maint: ^ how does that look for a nodepool release? | 20:51 |
| tobiash | lgtm | 20:52 |
| clarkb | double checkin | 20:52 |
| clarkb | sha1 and tag version lgtm | 20:53 |
| corvus | ok will push now | 20:56 |
| fungi | looks like it's pushed | 20:57 |
| fungi | (showed up in my git log when i did a remote update to check) | 20:57 |
| fungi | anyway, lgtm too | 20:57 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Scheduler ref-updated create/delete https://review.opendev.org/739198 | 20:59 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Get ref-updated modified files https://review.opendev.org/739078 | 20:59 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Update tests/base.py to use proper git data https://review.opendev.org/742746 | 20:59 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Add tests https://review.opendev.org/742747 | 20:59 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Reorg scheduler event process loop https://review.opendev.org/742748 | 20:59 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Allow ref-updated newrev+oldrev reconfiguration https://review.opendev.org/742749 | 20:59 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Merge upload logs modules into common role https://review.opendev.org/742732 | 21:07 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Consolidate common log upload code into module_utils https://review.opendev.org/742736 | 21:07 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Merge upload logs modules into common role https://review.opendev.org/742732 | 21:22 |
| openstackgerrit | Tobias Henkel proposed zuul/zuul-jobs master: Consolidate common log upload code into module_utils https://review.opendev.org/742736 | 21:22 |
| *** hamalq has joined #zuul | 21:27 | |
| *** nils has quit IRC | 22:05 | |
| *** harrymichal has quit IRC | 22:14 | |
| *** rlandy is now known as rlandy|bbl | 22:18 | |
| *** Goneri has quit IRC | 22:19 | |
| ianw | has anyone else noticed page-up/page-down not working on the new log view page? | 22:25 |
| clarkb | ya spacebar doesn't work either | 22:27 |
| clarkb | someone mentioend we may have to handle those events in the code directly (because react or patterfly?) | 22:27 |
| ianw | if i enable an event listener breakpoint on keyboard and push the pgup key, it hits. so that's something! | 22:27 |
| ianw | code: "PageUp" | 22:28 |
| ianw | so it even knows what key it was | 22:28 |
| ianw | then it breaks on "Toggle.tsx:63" | 22:29 |
| corvus | ianw: it also doesn't work on any page | 22:36 |
| ianw | that's good, i guess? :) | 22:37 |
| corvus | it's certainly not desirable :) but hopefully subject to a global fix | 22:37 |
| ianw | the only toggles mentioned in the code are handleKebabDropdownToggle | 22:38 |
| ianw | i'm interested to find out what a kebab dropdown toggle is | 22:38 |
| ianw | ok, it appears to be a dropdown menu that falls out of 3 stacked vertical dots. i guess that looks like meat on a stick? | 22:41 |
| clarkb | ha ya that image mostly works | 22:41 |
| fungi | and here i thought it was kibidango | 22:42 |
| fungi | or takoyaki | 22:43 |
| clarkb | swagger ui and react-dom both have entries for PageUp though they both seem to just be event integer id to name mappings | 22:44 |
| ianw | ok, so the breakpoints are react-dom.production.min.js -> kebabtoggle -> back to react-dom.production.min.js -> done | 22:45 |
| ianw | which makes me think clarkb's suggestion that something needs to register itself to this event is likely | 22:45 |
| ianw | it looks like react is maybe looping and sending the event to people who want to hear it | 22:45 |
| ianw | https://github.com/bvaughn/react-window/issues/46 | 22:48 |
| clarkb | there is a react-scroll-listener packages, though that seems pretty overkill | 22:48 |
| ianw | i think i saw references to "window" | 22:48 |
| ianw | i don't think we're using react-window though. but getting closer | 22:52 |
| clarkb | corvus: do we need an email for the nodepool release too? | 22:57 |
| *** tosky has quit IRC | 22:59 | |
| *** saneax has joined #zuul | 23:01 | |
| *** jamesmcarthur has joined #zuul | 23:04 | |
| jamesmcarthur | corvus: any issues/concerns with me tweeting the latest 3.19.1 security update? | 23:04 |
| corvus | clarkb: yes we do, jamesmcarthur: no concerns :) | 23:05 |
| jamesmcarthur | sweet! | 23:05 |
| jamesmcarthur | on it | 23:05 |
| corvus | clarkb: https://zuul.opendev.org/t/zuul/build/ae3d623963044d2a9619ac8be7461cb9 | 23:08 |
| corvus | didn't we fix that? | 23:08 |
| corvus | clarkb: https://zuul.opendev.org/t/zuul/build/ae3d623963044d2a9619ac8be7461cb9/console ran ensure-pip | 23:08 |
| clarkb | also how did zuul succeed if the fix was incomplete | 23:09 |
| clarkb | corvus: https://zuul.opendev.org/t/zuul/build/ae3d623963044d2a9619ac8be7461cb9/log/job-output.txt#356-357 did we skip that in the zuul case? | 23:09 |
| corvus | that deep link did not scroll in view | 23:10 |
| corvus | felixedel: i'm starting to suspect we have some serious scrolling-related problems with pf4 -- pageup/pagedown/arrows/spacebar don't work and neither does deep linking to lines in log files | 23:11 |
| clarkb | ah ok thats a specific case handler for suse so the skip is expected | 23:11 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Allow ref-updated newrev+oldrev reconfiguration https://review.opendev.org/742749 | 23:11 |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Get ref-updated modified files https://review.opendev.org/739078 | 23:11 |
| openstackgerrit | Ian Wienand proposed zuul/zuul master: [dnm] keypresses in log page https://review.opendev.org/742759 | 23:12 |
| clarkb | ianw: ^ you probably understand all of that better than we do. We added ensure-pip to opendev/base-jobs python releas job here https://opendev.org/opendev/base-jobs/commit/6e3f4db73886b281e0db6e25e4add91f41f5ac3f it ran in the job corvus linked but later import setuptools failed | 23:12 |
| clarkb | python verison mismatch maybe? | 23:12 |
| clarkb | corvus: ianw ya I think that may be it, we used python not `python3` | 23:14 |
| clarkb | https://zuul.opendev.org/t/zuul/build/ae3d623963044d2a9619ac8be7461cb9/console#2/0/3/ubuntu-bionic | 23:14 |
| clarkb | trying to find zuul case now to compare | 23:14 |
| ianw | yeah, that will not by default install python2 deps | 23:15 |
| clarkb | https://zuul.opendev.org/t/zuul/build/63f287ce85fe461d93b771a383cd57a9/console#3/0/3/ubuntu-bionic I think that is it | 23:15 |
| corvus | so why did that happen? | 23:15 |
| corvus | because the release job uses python by default? | 23:15 |
| corvus | why don't these roles work together? | 23:16 |
| clarkb | zuul sets release_python to python3 https://opendev.org/zuul/zuul/src/branch/master/.zuul.yaml#L214-L219 | 23:16 |
| clarkb | corvus: I think beacuse the release job started with python way back in the beginning of zuulv3 roles | 23:16 |
| clarkb | and changing that is potentially problematic? | 23:16 |
| *** jamesmcarthur has quit IRC | 23:16 | |
| corvus | whoops we changed it | 23:16 |
| clarkb | sure, but we changed it and it faield rather than building a bad python2 package with python3 ? | 23:17 |
| clarkb | (I don't know how big of a concern that is) | 23:17 |
| corvus | (i mean the problem would manifest as: "people using roles with default settings suddenly find it isn't working anymore" right?) | 23:17 |
| clarkb | corvus: yes, though with release artifacts that may result in "you released a bad artifact to pypi and now can't update it" | 23:17 |
| *** jamesmcarthur has joined #zuul | 23:17 | |
| clarkb | but I'm not sure how realistic that concern is | 23:17 |
| corvus | naturally we have changed it for check and gate but not release | 23:18 |
| openstackgerrit | James E. Blair proposed zuul/nodepool master: Use python3 for building releases. Everywhere. https://review.opendev.org/742761 | 23:21 |
| corvus | clarkb, ianw: ^ how about we use a thing we made to make sure that doesn't happen | 23:21 |
| clarkb | ++ may want to do the same in zuul too | 23:22 |
| corvus | clarkb: ahead of you :) | 23:22 |
| corvus | git review is um, setting up some suspense | 23:22 |
| openstackgerrit | James E. Blair proposed zuul/zuul master: Use project vars for release_python https://review.opendev.org/742762 | 23:23 |
| ianw | hrm at what point do with switch that in zuul-jobs? | 23:24 |
| corvus | ianw: maybe we should go aheand and start a zuul-discuss thread suggesting we do that real soon now and see if anyone has a reason not to? | 23:25 |
| clarkb | ianw: I think we can probably siwtch it there nowish, but should probably advertise the switch first | 23:25 |
| clarkb | corvus: ++ | 23:25 |
| *** jamesmcarthur has quit IRC | 23:26 | |
| ianw | yeah, it's just going to fail hard i guess on focal where no /usr/bin/python | 23:26 |
| ianw | and just work on distros where that is python3 already | 23:26 |
| *** jamesmcarthur has joined #zuul | 23:27 | |
| *** jamesmcarthur has quit IRC | 23:37 | |
| *** hamalq has quit IRC | 23:40 | |
| *** saneax has quit IRC | 23:43 | |
| openstackgerrit | Guillaume Chauvel proposed zuul/zuul master: Tenant reconfiguration: Get ref-updated modified files https://review.opendev.org/739078 | 23:43 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!