| *** gmann_pto is now known as gmann | 04:59 | |
| opendevreview | Dmitriy Rabotyagov proposed openstack/project-config master: Deprecate openstack-ansible-tests repository https://review.opendev.org/c/openstack/project-config/+/947629 | 06:24 |
|---|---|---|
| opendevreview | Dmitriy Rabotyagov proposed openstack/project-config master: Remove noop jobs for openstack-ansible-tests https://review.opendev.org/c/openstack/project-config/+/947665 | 12:38 |
| opendevreview | Karolina Kula proposed opendev/glean master: Add support for CentOS 10 keyfiles https://review.opendev.org/c/opendev/glean/+/941672 | 14:57 |
| clarkb | corvus: https://paste.opendev.org/show/bXJRMKf4tnCs5DcGbaSc/ more science. I think this shows that it very is likely just artifacting from the tracing system. Both sigint and sighup exhibit similar behaviors that change from attempt to attempt | 15:35 |
| clarkb | I didn't include kill -9 output because it is consistently boring with the shutdown callback not being called (as expected) | 15:35 |
| clarkb | I think I'm like 99% confident that this change from HUP to INT is safe now. I'll followup wtih gerrit discord and note the testing I did | 15:36 |
| clarkb | and left a detailed note on https://review.opendev.org/c/opendev/system-config/+/947540/ | 15:42 |
| clarkb | I've been totally nerd sniped. FOund https://www.brendangregg.com/Slides/dtrace_topics_java.pdf where brendan gregg has been doing this stuff with dtrace for decades | 15:51 |
| clarkb | its too bad that solaris died | 15:52 |
| clarkb | lots of great engineering and tooling in that system | 15:52 |
| mnaser | i'm sorry i keep bringing up weird git issues =) | 16:45 |
| mnaser | we use renovate to bump things to the latest to kee pthings updated and i ran into a weird issue | 16:45 |
| mnaser | what i can see the latest commit for zed-eol in https://opendev.org/openstack/heat-dashboard/src/tag/zed-eol/ is 5d1ef0faf10c77490051984e850ce0431054db7f | 16:45 |
| mnaser | however, somehow, renovate seems to think that's not the right one and attempts to change it to 6eda28f2ad0c74ccb5d9f0986811ff177a95f110 | 16:46 |
| mnaser | and when i put https://opendev.org/openstack/heat-dashboard/commit/6eda28f2ad0c74ccb5d9f0986811ff177a95f110 .. it looks like it's tehe same commit?! | 16:46 |
| mnaser | and then our docker build job fails because "fatal: git cat-file: could not get object info" | 16:46 |
| mnaser | whats also odd is that its not like this had multiple revs so maybe it was another rev or something.. https://review.opendev.org/c/openstack/heat-dashboard/+/857444 | 16:47 |
| clarkb | mnaser: I'm trying to confirm this is the case here (have to remembr the git magic) but there are two hashes associated with every tag. The first is the has hfor the tag object and the second is the hash for the thing the tag points to | 16:58 |
| clarkb | if you do git show zed-eol you get commit 5d1ef0faf10c77490051984e850ce0431054db7f so the commit itself is 5d1ef... now I'm just trying to confirm 6eda is the tag sha | 16:58 |
| clarkb | `git rev-parse zed-eol` says the hash for the tag is 6eda28f2ad0c74ccb5d9f0986811ff177a95f110 | 16:59 |
| mnaser | so i guess renovate uses rev-parse in this case, but buildkit checkouts doesnt like that | 17:00 |
| clarkb | one reason may be that you haven't fetched the tag | 17:00 |
| clarkb | the commit version may be present simply by being in the tree history. But the tag is a leaf | 17:01 |
| clarkb | so if you haven't fetched the tag properly then it wouldn't be there maybe | 17:01 |
| mnaser | we use the dockerfile ADD with git .. https://github.com/vexxhost/atmosphere/blob/stable/zed/images/horizon/Dockerfile#L28 it doesn't fetch much, we always have to do fetch unshallow afterwards to get pbr version to work properly | 17:02 |
| clarkb | unrelated I think I've decided to make difftastic my default diff tool for git | 17:11 |
| fungi | i'm not here, but `git show-ref ...` is what you use to get the tag object id hash | 18:12 |
| fungi | er, what i use anyway | 18:12 |
| clarkb | I thought there was a way to have git show show both of them too | 18:14 |
| clarkb | like tag xyz commit abc | 18:14 |
| clarkb | but I couldn't figure that out so maybe I'm dreaming it | 18:14 |
| fungi | i usually just use show and show-ref to get them separateely | 18:16 |
| *** dansmith_pto is now known as dansmith | 18:45 | |
| clarkb | I just remembered that we had paused noble image builds and the expected kernel fix should've arrived this week | 20:15 |
| * clarkb tries to hunt down if that happened | 20:15 | |
| clarkb | I believe the new 6.8.0-58.60 kernel is the expected fix | 20:24 |
| clarkb | based on the original bug and the bug it refers to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2104134 and https://bugs.launchpad.net/kernel-sru-workflow/+bug/2102529 | 20:24 |
| clarkb | I just rebooted review03 onto that kernel now ( Irealized getting it up to date before the move monday is a good idea) | 20:26 |
| clarkb | the project-config/nodepool/nodepool.yaml file says we are not pausing noble but noble dib-image-list says we are. I think that means the pause was requested on the command line | 20:32 |
| clarkb | timburke_: "configuration error: `tool.setuptools` must not contain {'ext-modules'} properties" is weird since https://setuptools.pypa.io/en/latest/userguide/ext_modules.html has a literal example doing just that | 20:46 |
| clarkb | but unfortunately I don't have any idea ideas about fixing it | 20:46 |
| fungi | i'm not here, but wrong setuptools version? | 20:48 |
| clarkb | ya thats my hunch at this point. Maybe the setuptools is too old to support that functionality | 20:50 |
| clarkb | looks like this build is happening in a docker image clled pyeclib-build-wheel-latest so maybe just a matter of updating that image | 20:51 |
| clarkb | and that seems to be based on the upstream manylinux images | 20:53 |
| clarkb | hrm the pyproject.toml files says it needs setuptools >=74.1 | 21:20 |
| clarkb | that is less than a year old but maybe ext modules support is even newer? | 21:21 |
| clarkb | 74.1 is the first versions that supports it according to the changelog | 21:24 |
| clarkb | I'm stumped without trying to reproduce this locally. I'll leave that for timburke_ | 21:24 |
| clarkb | oh its doing python setup.py bdist_wheel | 21:28 |
| clarkb | so maybe isn't invoking the >=74.1 installation | 21:28 |
| clarkb | timburke_: ya I think you need to update https://opendev.org/openstack/pyeclib/src/branch/master/pack_wheel.py#L104-L107 to use the build tool instead. That tools does seem to be installed in the manylinux wheels so you should be able to just use it | 21:33 |
| timburke_ | clarkb, thanks for looking at it! yeah, i wanted to try a thing, sorry if it set off any alarm bells for you ;-) | 23:57 |
| timburke_ | unfortunately, i don't think build is going to work for me (or at any rate, i couldn't figure out any way to have it spit out abi3 wheels) | 23:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!