| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 945907: WIP Make Gerrit Connection Replication Aware https://review.opendev.org/c/zuul/zuul/+/945907 | 02:21 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 946240: Always clear wake event of TPE event connectors https://review.opendev.org/c/zuul/zuul/+/946240 | 12:20 | |
| -@gerrit:opendev.org- Simon Westphahl proposed: [zuul/zuul] 946240: Always clear wake event of TPE event connectors https://review.opendev.org/c/zuul/zuul/+/946240 | 12:21 | |
| @f2ked:matrix.org | I'm trying to build the docker, I'm using docker compose and needed to set `PBR_VERSION`. Unfortunately, the resutling docker image did not install/copy-over `zuul/vendor` | 14:25 |
|---|---|---|
| ``` | ||
| File "/usr/local/lib/python3.11/site-packages/zuul/scheduler.py", line 28, in <module> | ||
| from zuul.vendor.contextlib import nullcontext | ||
| ModuleNotFoundError: No module named 'zuul.vendor' | ||
| ``` | ||
| @fungicide:matrix.org | f2ked: you need a working yarnpkg executable or it'll be skipped | 14:26 |
| @f2ked:matrix.org | fungi: where? the Dockerfile installs and runs it,right? | 14:28 |
| ``` | ||
| COPY web /tmp/src | ||
| # Explicitly run the Javascript build | ||
| RUN cd /tmp/src && yarn install -d && yarn build | ||
| ``` | ||
| @fungicide:matrix.org | it should, yes. maybe the problem is elsewhere, i had just noticed when building sdists that it was skipped if not present, i guess the container image builds have it as a hard dependency | 14:29 |
| @f2ked:matrix.org | I had to set `PBR_VERSION` to make pbr/sdists happy, I could not figure out why it was not happy with the git image (maybe because it is a submodule?) | 15:24 |
| @clarkb:matrix.org | if you are using submodules then you aren't using our dockerfile as I'm pretty sure therea re no submoduels there | 15:24 |
| @fungicide:matrix.org | yeah, i don't know how pbr would handle being run from within a submodule context either | 15:25 |
| @fungicide:matrix.org | we basically avoid submodules like the plague | 15:25 |
| @clarkb:matrix.org | it might expect the parent repo to respect its version schema and semantics | 15:25 |
| @f2ked:matrix.org | I'm using your Dockerfile with a small change and docker-compose | 15:26 |
| ``` | ||
| @f2ked:matrix.org | * I'm using your Dockerfile with a small change and docker-compose | 15:26 |
| ``` | ||
| build: | ||
| context: ./zuul | ||
| dockerfile: ../zuul.Dockerfile | ||
| ``` | ||
| @f2ked:matrix.org | the change merely set the UID of the zuul user to a required value | 15:27 |
| @clarkb:matrix.org | but where do submodules come into play? | 15:28 |
| @f2ked:matrix.org | but that is what `PBR_VERSION` is for.. but how do I get it to not skip `zuul/vendor` | 15:28 |
| @f2ked:matrix.org | I have a git repo with all my stuff and nodepool and zuul are submods | 15:28 |
| @f2ked:matrix.org | if I need to make them "side gits" then so be it | 15:29 |
| @fungicide:matrix.org | it's entirely possible that pbr's manifest generation is being confused by running in a submodule | 15:30 |
| @fungicide:matrix.org | normally pbr inspects git ls-files to decide what files should be packaged | 15:30 |
| @clarkb:matrix.org | and it generates version info from the git log (which it may also be confusing the parent repo log for zuul's log) | 15:31 |
| @clarkb:matrix.org | using PBR_VERSION short circuits that behavior and it will just use the version you set in that var | 15:31 |
| @clarkb:matrix.org | usually that version is expected to be used by distribution packagers who may be carrying local patches to their packages that they need udpated versions for | 15:31 |
| @clarkb:matrix.org | you probably need to read a more complete log of the image build to see where vendor/contextlib.py is getting lost. Maybe even inspect each stage of the build process to see where it disappears | 15:35 |
| @f2ked:matrix.org | when in the submodule directory.. it looks like a normal repo.. but it may be confused the `.git` is (now by default) a file reference to a dir in the parent `.git`. maybe if I force it to have it's own `.git` it might work? | 15:35 |
| @clarkb:matrix.org | side note we support pythons that are all new enough taht we no longer need to vendor that code and can rely on stdlib | 15:35 |
| @f2ked:matrix.org | I've scoured the logs.. it simply isn't mentioned | 15:36 |
| @clarkb:matrix.org | ok I think what is happening is `COPY . /tmp/src` line 60 of the Dockerfile gets your submodule without the git info. Then pbr can't do its job | 15:38 |
| @clarkb:matrix.org | it can't determine a version and it can't determine which files to incldue dynamically because you are no longer providing the git info | 15:38 |
| @fungicide:matrix.org | ah, yeah i was confusing the vendored python libs with the vendored javascript libs for the dashboard, but i guess those end up in a different directory entirely | 15:40 |
| @f2ked:matrix.org | ok.. but the version issue is solved by `PBR_VERSION`, is there anything I can do to get the `vendor` stuff copied over? | 16:30 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/zuul] 945907: WIP Make Gerrit Connection Replication Aware https://review.opendev.org/c/zuul/zuul/+/945907 | 16:38 | |
| @fungicide:matrix.org | you might be able to manually construct the python packaging manifest to include those files rather than relying on pbr to find them through git inspection | 16:44 |
| @clarkb:matrix.org | corvus: I see the stack based on eatmydata is approved (no concerns there from me). I do wonder if maybe we should update ensure-zookeeper to have a use eatmydata option in addition to using tmpfs though | 16:45 |
| @clarkb:matrix.org | thoguh maybe its utility for nodepool is quickly going away | 16:46 |
| @jim:acmegating.com | yeah if we feel like that would be a good general option, i think it would be okay to add. would simplify the test setup a bit (get rid of an extra restart). just didn't want to commit zuul-jobs to it until we knew it was worthwhile | 16:47 |
| @jim:acmegating.com | i'm not sure i'd bother changing the nodepool jobs. could cause problems with no benefit. | 16:48 |
| @clarkb:matrix.org | ya and nodepool is winding down so unlikely to get major changes that need to be more memory nible | 16:49 |
| @clarkb:matrix.org | * ya and nodepool is winding down so unlikely to get major changes that need to be more memory nimble | 16:49 |
| @clarkb:matrix.org | zuulians there is discussion on the gerrit mailing list about updating the zuul results summary plugin that ianw wrote to be a checks plugin. I think having a zuul checks plugin would be great either way but wanted to make note of that here in case anyone other than opendev is using that plugin and may need to pin an older version while changes occur | 16:51 |
| @clarkb:matrix.org | its still a proposal and not necessarily going to happen but you should be aware if using that plugin | 16:52 |
| @jim:acmegating.com | i think we should request that they not gut and replace the existing plugin | 16:57 |
| @clarkb:matrix.org | corvus: do you want to respond or should I? I'm not sure if you are subscribed | 16:58 |
| @jim:acmegating.com | there are a lot of open questions about how a checks plugin would actually integrate with various types of zuul instances; also the minor fact that zuul makes no public api guarantees. it would be good to retain our existing codebase while that settles out. | 16:59 |
| @jim:acmegating.com | i will reply | 16:59 |
| @jim:acmegating.com | done, hope that captures things | 17:15 |
| @clarkb:matrix.org | I see it and looks good to me | 17:19 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul-jobs] 946289: Add eatmydata support to ensure-zookeeper https://review.opendev.org/c/zuul/zuul-jobs/+/946289 | 17:40 | |
| @f2ked:matrix.org | fungi: thanks for the tip.. FWIW this solves it for me | 17:43 |
| ``` | ||
| diff --git a/MANIFEST.in b/MANIFEST.in | ||
| index e571dfa8b..8a3ee29b4 100644 | ||
| --- a/MANIFEST.in | ||
| +++ b/MANIFEST.in | ||
| @@ -4,6 +4,7 @@ include zuul/ansible/*/* | ||
| include zuul/ansible/*/*/* | ||
| include zuul/driver/github/graphql/* | ||
| include zuul/lib/ansible-config.conf | ||
| +include zuul/vendor/* | ||
| include zuul/web/static/* | ||
| include zuul/web/static/static/*/* | ||
| ``` | ||
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 946176: Fix change cache test races https://review.opendev.org/c/zuul/zuul/+/946176 | 17:52 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 946166: Use eatmydata with zookeeper in tests https://review.opendev.org/c/zuul/zuul/+/946166 | 17:53 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 945743: Remove tenant references from project parsing https://review.opendev.org/c/zuul/zuul/+/945743 | 18:25 | |
| @clarkb:matrix.org | corvus: I think the stack behind ^ needs a recheck | 18:37 |
| @jim:acmegating.com | yep was just waiting for the timeout. i checked this morning and it looked like we're at 7% failure rate. which is... not bad compared to recent history :) | 18:57 |
| @clarkb:matrix.org | things definitely seem happier | 18:58 |
| @jim:acmegating.com | rare opportunity to see the active window in action on the zuul tenant :) | 19:02 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul-jobs] 946289: Add eatmydata support to ensure-zookeeper https://review.opendev.org/c/zuul/zuul-jobs/+/946289 | 19:05 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 20:21 | |
| - [zuul/zuul-jobs] 946289: Add eatmydata support to ensure-zookeeper https://review.opendev.org/c/zuul/zuul-jobs/+/946289 | ||
| - [zuul/zuul-jobs] 946304: Remove Red Hat support from ensure-zookeeper https://review.opendev.org/c/zuul/zuul-jobs/+/946304 | ||
| @jim:acmegating.com | red hatters -- i'd like feedback on what to do with that role in 946304 (remove rh support? or maybe there's a way to fix testing? or leave it alone and add an "untested, best effort" caveat to the docs?) | 20:22 |
| @clarkb:matrix.org | is there a link to a failure? that may be helpful in understanding if fixing it is reasonable | 20:25 |
| @jim:acmegating.com | https://zuul.opendev.org/t/zuul/build/e0864716d7ef4aec8ad0e0443bc8ffa2 | 20:26 |
| @jim:acmegating.com | i just added that as a comment to the change too | 20:26 |
| @jim:acmegating.com | that's on the change above it, but not running with the eatmydata package | 20:26 |
| @jim:acmegating.com | (so effectively the current state) | 20:28 |
| @clarkb:matrix.org | looking at https://mirror.stream.centos.org/9-stream/AppStream/x86_64/os/Packages/ I thik they may have dropped the latest jdk packge and now you have to use something like java-21-openjdk-headless | 20:30 |
| @clarkb:matrix.org | 17, 11, and 8 are also available | 20:30 |
| @clarkb:matrix.org | (though 8 is 1.8.0) | 20:30 |
| @clarkb:matrix.org | but then I'm not sure if you get a generic java binary you can run so probably best if someone who cares digs in | 20:34 |
| @clarkb:matrix.org | oh I guess zookeeper's startup script may find java for itself. Maybe that is all that needs to change | 20:34 |
| @jim:acmegating.com | yeah if we have to keep bumping those versions... and there's no one interested in bumping those versions... then we should probably retire it. | 21:17 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 946308: Make test_zuul_query_ltime more deterministic https://review.opendev.org/c/zuul/zuul/+/946308 | 21:49 | |
| @iwienand:matrix.org | > updating the zuul results summary plugin that ianw wrote to be a checks plugin | 22:48 |
| that sounds great, it was just too much time for me to figure out all the polymer and shadow doms and whatever to get it done right :) | ||
| @fungicide:matrix.org | to be clear, you don't want to make a mistake drawing the binding sigil for those before you summon | 22:50 |
| @iwienand:matrix.org | 🤣 i _might_ sign on the dotted line if I had any indication that the whole framework wouldn't be declared as deprecated and the completely incompatible V3 would be released 5 minutes later | 22:53 |
| @iwienand:matrix.org | (grumpy old person who regularly works with .c files with (c) dates in the 80's :) | 22:54 |
| @jim:acmegating.com | to be clear, i think it sounds great too -- but on the mailing list i'm pushing that they do that in a new repo :) | 22:55 |
| @fungicide:matrix.org | that sums up my impression of web frameworks to a t | 22:55 |
| @jim:acmegating.com | they're also not really talking about updating it to use the checks api -- they're talking about a completely different plugin codebase | 22:56 |
| @jim:acmegating.com | like, the existing plugin reads the data from the review messages, and writes it to the web interface in a tab | 22:57 |
| @jim:acmegating.com | the code base they are talking about gets the data from the zuul api, and writes it to the checks api | 22:57 |
| @jim:acmegating.com | so there's almost nothing in common | 22:57 |
| @jim:acmegating.com | scraping from the review message and writing to the checks api might be something that would make sense to do in that repo, but i don't think that's what's on offer. | 23:02 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 945744: Stop copying ProjectConfig objects when adding to layout https://review.opendev.org/c/zuul/zuul/+/945744 | 23:10 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 945745: Check for nodeset label perms in layout https://review.opendev.org/c/zuul/zuul/+/945745 | 23:15 | |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: | 23:28 | |
| - [zuul/zuul] 945746: Check for job timeouts in layout https://review.opendev.org/c/zuul/zuul/+/945746 | ||
| - [zuul/zuul] 945747: Resolve required-projects in layout https://review.opendev.org/c/zuul/zuul/+/945747 | ||
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 945748: Resolve include-vars projects in layout https://review.opendev.org/c/zuul/zuul/+/945748 | 23:28 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 946315: Fix change cache test races more https://review.opendev.org/c/zuul/zuul/+/946315 | 23:56 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!