| mordred | oh fun ^^ | 00:12 |
|---|---|---|
| mordred | would be nice to be able to completely retire that PPA | 00:13 |
| mordred | clarkb: I did just add support for python-version files should you decide to start pyenv-ing in places | 00:14 |
| mordred | :) | 00:14 |
| tonyb | Yeah RHEL-8 is python 3.8 so perhaps that will stick around. | 00:17 |
| tonyb | mordred: I *think* for AFS clients we can switch to kAFS which I think would help with retiring that PPA. | 00:18 |
| tonyb | I kinda have it in my game plan to treat Noble as a switchover point. | 00:19 |
| mordred | ++ | 00:19 |
| tonyb | I need to verify what is available/works if you want the AFS utils bos, vos etc etc | 00:19 |
| tonyb | oooo /me remembered ianw's https://review.opendev.org/c/opendev/system-config/+/623974 patches | 00:21 |
| corvus | we should see what the oldest patch we can merge in 2024 is | 00:22 |
| mordred | corvus: https://review.opendev.org/c/opendev/ansible-role-puppet/+/592221 | 00:24 |
| tonyb | :) | 00:25 |
| mordred | actually - https://review.opendev.org/c/opendev/system-config/+/418469 is the oldest I've got in my outbound that isn't in merge conflict and also isn't for an abandoned project (since that's cheating) | 00:26 |
| opendevreview | James E. Blair proposed ttygroup/gertty master: Allow the configured git-url to override gerrit https://review.opendev.org/c/ttygroup/gertty/+/169810 | 00:28 |
| corvus | honestly, i have no idea if that change makes any sense any more, and that's obviously cheating too. :) | 00:29 |
| mordred | https://review.opendev.org/c/x/entropy/+/119920 is my oldest without a merge conflict :) | 00:32 |
| mordred | wow - gertty is 9 years old! | 00:33 |
| mordred | corvus: here's one of yours: https://review.opendev.org/c/opendev/log_processor/+/254312 | 00:33 |
| mordred | corvus: I think that gertty patch might win for "old and also passes tests after a recheck" | 00:37 |
| fungi | i have a couple of open changes from 2014 but they're in abandoned repos | 00:42 |
| mordred | yah | 00:50 |
| fungi | also i have a bunch of old puppet-related changes i should get around to abandoning | 00:54 |
| fungi | but not tonight | 00:54 |
| opendevreview | Merged opendev/system-config master: Test mirror services on noble https://review.opendev.org/c/opendev/system-config/+/921771 | 01:01 |
| opendevreview | Tony Breeds proposed opendev/system-config master: Add an opendev specific build of mediawiki https://review.opendev.org/c/opendev/system-config/+/921321 | 04:49 |
| opendevreview | Tony Breeds proposed opendev/system-config master: DNM: Initial dump or mediawiki role and config https://review.opendev.org/c/opendev/system-config/+/921322 | 04:49 |
| opendevreview | Adrian Vladu proposed openstack/project-config master: WOIP: x/cloudbase-init: move to GitHub https://review.opendev.org/c/openstack/project-config/+/923044 | 12:18 |
| opendevreview | Adrian Vladu proposed openstack/project-config master: WOIP: x/cloudbase-init: move to GitHub https://review.opendev.org/c/openstack/project-config/+/923044 | 12:24 |
| opendevreview | Julia Kreger proposed openstack/diskimage-builder master: bootloader: Strip prior console settings https://review.opendev.org/c/openstack/diskimage-builder/+/922961 | 13:59 |
| *** dmellado07553 is now known as dmellado0755 | 14:51 | |
| opendevreview | Julia Kreger proposed openstack/diskimage-builder master: bootloader: Strip prior console settings https://review.opendev.org/c/openstack/diskimage-builder/+/922961 | 15:16 |
| mnaser | I started building a bunch of elements for Cluster API to get kubernetes cluster images going, and I have successfully did it and its working really well with Zuul publishing them | 16:08 |
| mnaser | However, my only challenge is building rocky images, it seems that the only path is `rocky-container` ? but it seems for some reason im getting some CA issues when pulling certain things inside rocky, is there a "use cloud image" option for Rocky by any chance ? | 16:08 |
| clarkb | mnaser: are you specifically asking about using disk image builder? If so no. | 16:13 |
| mnaser | clarkb: yeah, dib https://github.com/vexxhost/capo-image-elements -- rocky-container is failing because for some reason containerd is getting tis cert verification errors | 16:13 |
| clarkb | we successfully build rocky images using that element though and don't have CA issues as far as I know. NeilHanlon also hangs out around here and may be able to help debug | 16:13 |
| mnaser | I suspect maybe its a container issue then | 16:13 |
| mnaser | containerd** | 16:13 |
| clarkb | mnaser: https://nb02.opendev.org/rockylinux-9-eee1793153fc402fbcf297dc68d96e1e.log I think that is our most recent build if you want to compare logs | 16:19 |
| mnaser | clarkb: thanks for sharing that, I think this is a containerd thing where its probably looking for the wrong path for the CA on rocky | 16:31 |
| mnaser | because I just realized I do a few curl's in some of the elements that work just fine | 16:31 |
| clarkb | corvus: sean-k-mooney pointed out an interesting zuul reporting result here: https://review.opendev.org/c/openstack/nova/+/915735 all of the builds were cancelled and it reported -2 from the gate | 17:32 |
| clarkb | corvus: looking at logs on zuul01 it appears that 915735's parent (915734) had a failing job causing 915735 to have its jobs cancelled. I would've expected 915735 to get reenqueued with new jobs and not reported teh acncelled state | 17:32 |
| clarkb | however it seems that 915734 gets reenqeueud instead at some later point and merges but 915735 never does and we report the cancelled state | 17:33 |
| clarkb | corvus: https://zuul.opendev.org/t/openstack/builds?change=915734&skip=0 I think the RETRY for nova-tox-validate-backport here is triggering a gate reset that never completes because retry isn't a true failing state. That then effectively orphans the reset for the child change (915735) | 17:39 |
| clarkb | I'll move this conversation over to the zuul matrix room now that I think I've done enough debugging on the opendev side | 17:39 |
| mnaser | is there a way to make DIB "break" its build process and let me chroot into its mounts to reproduce/test things? | 19:04 |
| fungi | i'd be surprised if there's not a command-line option that tells it not to clean up after a failure | 19:08 |
| clarkb | mnaser: fungi: you edit the element script and put a line with bash in it | 19:11 |
| mnaser | oh wow that's so straightforward lol | 19:11 |
| mnaser | good call | 19:11 |
| clarkb | or sh or whatever shell is in the chroot and it should basically stop at that point in time giving you full interactive access then when you exit it will proceed | 19:11 |
| fungi | fair enough. injecting an interactive bash invocation in the middle of a script seems like a hackish option compared to a --keep or --no-cleanup in the cli, but i agree that's fairly straightforward | 19:13 |
| fungi | and allows you to pick the exact breakpoint | 19:14 |
| mnaser | honestly even if that option existed if I knew about bash thats how I'd do it | 19:14 |
| fungi | or breakpoints if you want more than one, and then continue the process by exiting | 19:14 |
| mnaser | I'm more of a "echo XXXXXXXXXXXXX" debug type of person :p | 19:14 |
| fungi | yeah, i agree that's more flexible | 19:14 |
| fungi | i definitely do my fair share of `echo $somevarimwonderingabout` or `print("this is the step we're executing")` | 19:16 |
| mnaser | yeah old habits die hard, I'd struggle using a proper debugger | 19:16 |
| mnaser | but in other cool news we went from using packer inside k8s to a bunch of lightweight disk image builder elements and the image build process went from an unreliable 20-30 minutes to a reliable 8 minutes and half the size of the final image | 19:17 |
| fungi | hah! take that, dib detractors! | 19:18 |
| mnaser | interesting | 19:29 |
| mnaser | env | grep /usr/lib/ssl/certs/ca-certificates.crt => SSL_CERT_FILE=/usr/lib/ssl/certs/ca-certificates.crt | 19:30 |
| mnaser | stat: cannot statx '/usr/lib/ssl/certs/ca-certificates.crt': No such file or directory | 19:30 |
| mnaser | thats my issue, time to find out where thats coming ffrom | 19:30 |
| mnaser | aaaand absolutely no reference to that anywhere in dib | 19:31 |
| mnaser | ok it's slurping that from my existing environment variables when im running it | 19:32 |
| fungi | mnaser: i have a /usr/lib/ssl/certs/ca-certificates.crt on my debian systems but no package installed which claims that file, so it's presumably auto-generated by some post-install step | 19:35 |
| mnaser | fungi: yeah but I think what's happening is that that SSL_CERT_FILE env propagates into my rocky build, and that file doesnt exist on rocky | 19:36 |
| fungi | ls that file supposed to be on the host system or in the chroot? | 19:36 |
| fungi | aha@ | 19:36 |
| mnaser | so I am getting tls issues because containerd is using that env variable | 19:36 |
| fungi | yeah that would make sense. maybe the environment isn't being properly sanitized or reset | 19:36 |
| mnaser | thats odd because im running this through Zuul and seeing the same issue, so maybe something in ubuntu that adds it | 19:36 |
| mnaser | or maybe it's something else, I just tried to unset it and retry locally | 19:37 |
| mnaser | ok, successfully built locally, running remotely now and adding `env` and I'll see | 19:40 |
| mordred | mnaser: you know - the funny thing is - people love to detract and then point to things like packer - and I'm always like "yeah, we did the packer-style thing back in the early 2010's - there is a reason we stopped ... it's slow and flaky and always will be" :) | 20:52 |
| opendevreview | Ghanshyam proposed openstack/project-config master: Retire kuryr-kubernetes and tempest plugin: end gate and update acl https://review.opendev.org/c/openstack/project-config/+/923072 | 21:38 |
| fungi | but, but... packer! | 21:54 |
| opendevreview | James E. Blair proposed zuul/zuul-jobs master: Install the docker-compose-plugin when installing upstream docker https://review.opendev.org/c/zuul/zuul-jobs/+/923083 | 23:26 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!