| jrosser_ | noonedeadpunk: are you around this week? | 07:54 |
|---|---|---|
| noonedeadpunk | hey, yep | 08:20 |
| jrosser_ | noonedeadpunk: just got meetings for a while right now but I did a bunch of work on the unmaintained branches and got to the point where I need to check things with you | 09:29 |
| jrosser_ | looks like unmaintained/<blah> is not necessaritly present for all components, (magnum, vs. magnum-ui i think is one example) | 09:30 |
| jrosser_ | kind of messy somehow | 09:30 |
| noonedeadpunk | jrosser_: I assume then <blah>-eol tag should be present, which next step after unmaintained phase | 09:34 |
| noonedeadpunk | as each project is free to move from unmaintained to eol state anytime | 09:35 |
| noonedeadpunk | but seems that magnum has skipped unmaintained at all | 09:36 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible unmaintained/victoria: DNM - testing unmaintained branch https://review.opendev.org/c/openstack/openstack-ansible/+/916175 | 09:44 |
| jrosser_ | maybe for unmaintained branches it is best to work on them one at a time (i did go slightly out of my mind trying to fix 3 at once) | 10:48 |
| jrosser_ | victoria is the oldest and the most close to working, this will show the next lot of problems which is tempest plugin versions being too new https://review.opendev.org/c/openstack/openstack-ansible/+/916175 | 10:48 |
| noonedeadpunk | yeah, sunsetting 3 branches at once was a bit /o\ ofc | 10:57 |
| noonedeadpunk | I wonder if Xena tempest versions might be "fine" for Victoria | 10:58 |
| noonedeadpunk | as we laready have SHAs pinned for X: https://opendev.org/openstack/openstack-ansible/src/branch/unmaintained/xena/playbooks/defaults/repo_packages/openstack_testing.yml | 10:58 |
| noonedeadpunk | But it might be really tough to find proper ones for V otherwise | 10:59 |
| jrosser_ | i think i fixed one somewhere by looking through the pypi release dates for versions released at the same time as V | 11:08 |
| jrosser_ | so that could be an approach | 11:09 |
| jrosser_ | perhaps we do miss something here when we make a release by not pinning down the tempest plugin versions to whatever is current, but I don't know if that is valid | 11:09 |
| KSR-DK | Hello all. I'm trying to reconfigure HAproxy setting for horizon, after running the upgrade to 2023.1. I've deleted the horizon file in /etc/haproxy/conf.d and have executed openstack-ansible playbooks/haproxy-install.yml, openstack-ansible playbooks/setup-openstack --tags install/configure - which has left my haproxy config without a horizon backend.. Any advice? | 12:19 |
| jrosser_ | KSR-DK: it is very useful to check the release notes when you do an upgrade, see https://docs.openstack.org/releasenotes/openstack-ansible/2023.1.html | 12:21 |
| jrosser_ | there you will find "Haproxy services are now configured separately at the beginning of each service playbook." | 12:21 |
| jrosser_ | so there is a major change with the way the haproxy config is done starting in 2023.1, it will be done for each service by the playbook that deploys that service, not like before where the haproxy-install playbook did the config for all services in one step | 12:22 |
| jrosser_ | i would also say that it should never be necessary to delete the file from /etc/haproxy/conf.d, the way ansible works will ensure that the content you want is placed there automatically | 12:23 |
| KSR-DK | Thanks jrosser - I misswrote.. I meant playbooks/os_horizon --tags install/configure, and yeah I noticed. I ended up doing the openstack-ansible playbooks/haproxy-install.yml as well out of pure frustration ;-) | 12:25 |
| KSR-DK | I removed the file in conf.d to give myself aneasy check on wether the file was placed there.. | 12:27 |
| KSR-DK | Maybe I need to run os_horizon with no tags? | 12:28 |
| andrewbonney | You may need these tags: --tags haproxy-service-config | 12:28 |
| KSR-DK | THANKS andrewbonney - that did the trick. | 12:35 |
| KSR-DK | Did I mis that in the DOCS somewhere? | 12:35 |
| andrewbonney | It's covered in the distribution upgrade guide, but generally for tags you really need to browse the Ansible code as it's hard to exhaustively document and keep up to date | 12:37 |
| andrewbonney | https://docs.openstack.org/openstack-ansible/latest/admin/upgrades/distribution-upgrades.html | 12:37 |
| KSR-DK | andrewbonney - thanks, It did not cross my mind to look in the Distro upgrade guide for this, since I thought it would either be a deployment config thing, or a specific bullet the Operations Guide. Annyway - THANKS! | 12:46 |
| noonedeadpunk | huh, seems I found some reall weirdness in bootstraping process regarding collections.... | 13:12 |
| noonedeadpunk | like git+file:// is being replaced by // and is considered as zuul thing | 13:12 |
| noonedeadpunk | here https://opendev.org/openstack/openstack-ansible/src/branch/stable/2023.2/scripts/get-ansible-collection-requirements.yml#L59-L67 | 13:12 |
| opendevreview | Merged openstack/openstack-ansible-ops master: Correct supported release for mcapi_vexxhost https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916536 | 13:16 |
| jrosser_ | noonedeadpunk: here probably? | 13:18 |
| jrosser_ | https://opendev.org/openstack/openstack-ansible/src/branch/stable/2023.2/scripts/get-ansible-collection-requirements.yml#L52 | 13:18 |
| jrosser_ | zuul_src_path is not defined i guess, could be '' | 13:18 |
| jrosser_ | the only time that things should be "adjusted for CI" is when that stat task thinks it has found the repo on disk | 13:19 |
| noonedeadpunk | yeah, exactly | 13:28 |
| noonedeadpunk | but then you don't have CI, but have "git+file" which still ends up as repo on disk and thus adjusted | 13:28 |
| jrosser_ | i expect i have not tested git+file:// here | 13:29 |
| noonedeadpunk | Just trying 2 things, not surte what's better | 13:29 |
| jrosser_ | unfortuantely we can't add a when to that task | 13:30 |
| jrosser_ | oh well, actually not sure | 13:30 |
| noonedeadpunk | can't we? | 13:30 |
| jrosser_ | you see the variables a terribly badly named here | 13:30 |
| noonedeadpunk | so one way would be to add condition here: https://opendev.org/openstack/openstack-ansible/src/branch/stable/2023.2/scripts/get-ansible-collection-requirements.yml#L66 | 13:31 |
| jrosser_ | `zuul_collections` needs to have all the collections regardless if they are actually zuul repos or not | 13:31 |
| noonedeadpunk | for the ternary | 13:31 |
| noonedeadpunk | and another would be block and another ternary | 13:31 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Better check if we're in CI during collection bootstrap https://review.opendev.org/c/openstack/openstack-ansible/+/916633 | 13:32 |
| noonedeadpunk | ie ^ (will adjust commit msg now) | 13:32 |
| jrosser_ | the only thing i wondered is that galaxy_collections_list_ci is not defined in the ternary outside ci | 13:35 |
| noonedeadpunk | ugh, yes | 13:36 |
| noonedeadpunk | and this also doesn't fully protect... | 13:37 |
| noonedeadpunk | though it somehow passed locally | 13:37 |
| opendevreview | Merged openstack/openstack-ansible-ops master: Use upstream version of kubernetes collection. https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916581 | 13:39 |
| noonedeadpunk | but actually ansible acts slightly weirdly with dir, as it still tries to reach galaxy - basically that's where things failed for me, as env doesn't have even proxy | 13:40 |
| noonedeadpunk | it's completel air-gapped for outgoing connections | 13:40 |
| jrosser_ | does it fail if it can't reach galaxy? | 13:44 |
| noonedeadpunk | yup | 13:45 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Better check if we're in CI during collection bootstrap https://review.opendev.org/c/openstack/openstack-ansible/+/916633 | 13:46 |
| noonedeadpunk | and takes ages as well | 13:46 |
| noonedeadpunk | damiandabrowski: fwiw ^ | 13:47 |
| jrosser_ | that really is sad - i wonder if we could add another code path that just used git clone for that instead of galaxy | 13:49 |
| jrosser_ | because really thats pretty much how we use ansible-galaxy at the moment when getting everything from github | 13:50 |
| noonedeadpunk | so, when you do git - it's not needeing galaxy | 13:50 |
| jrosser_ | yes exactly | 13:50 |
| noonedeadpunk | but somehow when you do type: dir - it does | 13:50 |
| jrosser_ | oooooh right i see | 13:51 |
| noonedeadpunk | (at least it seems that it does as that's why I become looking there) | 13:51 |
| jrosser_ | well - gitweb + apache is really not very much work | 13:51 |
| noonedeadpunk | and then bootstrap script does substitude your git_file to dir and it's weird | 13:51 |
| jrosser_ | and that would give you a local place to git clone from with http | 13:52 |
| noonedeadpunk | I"m doing very stupid thing lol | 13:52 |
| opendevreview | Merged openstack/openstack-ansible-ops master: Fix documentation for example haproxy setup https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916583 | 13:52 |
| noonedeadpunk | just trying to avoid gitlab auth by osa bootstrap script, by placing requirements to the deploy host with zuul | 13:52 |
| jrosser_ | that does feel like an ansible-galaxy bug then | 13:53 |
| noonedeadpunk | probably, it would be easier to auth after all... | 13:53 |
| noonedeadpunk | yeah, I guess I need to investigate that further, but still we're doing quite unexpected thing right now | 13:54 |
| noonedeadpunk | where you don't get version you would expect from git as it's treated as dir... | 13:54 |
| noonedeadpunk | anyway | 13:54 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Better check if we're in CI during collection bootstrap https://review.opendev.org/c/openstack/openstack-ansible/+/916633 | 13:55 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Add CAPI job to OPS repo https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916640 | 14:02 |
| opendevreview | Merged openstack/openstack-ansible-ops master: Improve magnum cluster api documenation https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916584 | 14:05 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_nova stable/2023.2: Ensure TLS is enabled properly for cell0 mapping DB connection https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/916453 | 14:12 |
| opendevreview | Merged openstack/openstack-ansible-ops master: Add docker-image-py requirement to documentation https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916588 | 14:17 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_magnum master: Allow zuul job variables to be inserted into user_variables https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/916647 | 15:42 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Test all supported versions of k8s workload cluster with magnum-cluster-api https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916649 | 16:00 |
| opendevreview | Merged openstack/openstack-ansible-os_magnum master: Add job to test Vexxhost cluster API driver https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/905199 | 16:05 |
| jrosser_ | \o/ | 16:14 |
| opendevreview | Merged openstack/openstack-ansible-os_neutron stable/2023.2: Fix multiline yaml formatting in neutron systemd services https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/915948 | 16:16 |
| opendevreview | Merged openstack/openstack-ansible-os_horizon master: Add support for deploying Horizon with Skyline https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/913439 | 16:22 |
| noonedeadpunk | sweeet | 16:26 |
| noonedeadpunk | wonder why https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916640?tab=change-view-tab-header-zuul-results-summary doesn't work though | 16:28 |
| noonedeadpunk | `internal endpoint for container-infrastructure-management service in RegionOne region not found` | 16:29 |
| noonedeadpunk | `container-infrastructure-management` looks weird? | 16:29 |
| jrosser_ | errrr | 16:32 |
| jrosser_ | oh is that something trying to look up the magnum endpoint in the service catalog? | 16:33 |
| noonedeadpunk | I think it is | 16:33 |
| jrosser_ | oh | 16:34 |
| jrosser_ | `+ export SCENARIO=aio_lxc` | 16:34 |
| jrosser_ | that'll be why | 16:34 |
| noonedeadpunk | huh | 16:34 |
| noonedeadpunk | ok | 16:34 |
| jrosser_ | not sure we have that much freedom with the job name | 16:35 |
| noonedeadpunk | it explains that | 16:35 |
| jrosser_ | or the string splitting is incorrect somehow | 16:35 |
| noonedeadpunk | I think you're right about freedom | 16:35 |
| noonedeadpunk | Though I was really confused about vexxhost.kubernetes.kubernetes being run along with k8s container being created | 16:36 |
| noonedeadpunk | export ACTION=ops | 16:36 |
| jrosser_ | oh well thats part of the horribleness of this whole setup | 16:37 |
| jrosser_ | all these files get dumped into place in the job pre playbook https://github.com/openstack/openstack-ansible-ops/tree/master/mcapi_vexxhost/playbooks/files/openstack_deploy | 16:38 |
| noonedeadpunk | ah, ok | 16:38 |
| noonedeadpunk | fair | 16:38 |
| jrosser_ | which kind of does some of what SCENARIO would do, but differently | 16:38 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Add CAPI job to OPS repo https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916640 | 16:39 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Test all supported versions of k8s workload cluster with magnum-cluster-api https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916649 | 16:42 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Test all supported versions of k8s workload cluster with magnum-cluster-api https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916649 | 16:43 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Test all supported versions of k8s workload cluster with magnum-cluster-api https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916649 | 16:44 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-ops master: Test all supported versions of k8s workload cluster with magnum-cluster-api https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916649 | 16:51 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Remove openstack-ansible-backup reference https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916650 | 16:52 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Remove readme include from repo docs https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916651 | 17:14 |
| jrosser_ | btw if you have any good ideas on how to publish docs for the cluster api subdirectory that would be nice | 17:16 |
| jrosser_ | i did a bunch of reading on sphinx and it seems that it's not really intended for "monorepo" with lots of pretty independant subdirs | 17:16 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Add documentation for ELK to the renderred docs https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916652 | 17:17 |
| noonedeadpunk | it really is not | 17:17 |
| noonedeadpunk | but that should work | 17:17 |
| noonedeadpunk | a bit nasty/hacky though... | 17:18 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Add ClusterAPI documentation to renderred docs https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916653 | 17:23 |
| jrosser_ | we maybe wanting a symlink there becasue those docs are already in a "known" location (see the magnum-cluster-api slack for example) | 17:26 |
| jrosser_ | perhaps that doesnt work - i dont know | 17:27 |
| noonedeadpunk | ah, you mean that moving README.rst is not right? | 17:28 |
| noonedeadpunk | I just confused by having 2 readme files.... | 17:28 |
| noonedeadpunk | So thought that might be it's not too late to move... | 17:28 |
| jrosser_ | that codes been merged for kind of months now and the location of the doc is linked elsewhere | 17:29 |
| jrosser_ | but we can change that if we want, but should be mindful of breaking links to docs | 17:30 |
| ThiagoCMC | Hey folks! Is OSA master branch good to test with Ubuntu 22.04, Caracal and Ceph 19? | 17:31 |
| noonedeadpunk | we need to land couple of things and probably cut a beta release once it's done... | 17:31 |
| noonedeadpunk | but sure, you can try that out :) | 17:32 |
| ThiagoCMC | Cool! :-D | 17:32 |
| noonedeadpunk | ThiagoCMC: though, I don't think Ubuntu has already published packages for 2024.1? | 17:32 |
| ThiagoCMC | Yep, already there | 17:32 |
| * noonedeadpunk just recalled you do distro path | 17:32 | |
| noonedeadpunk | ok, we haven't updated for that | 17:32 |
| ThiagoCMC | No worries... Count on me to test it extensively. | 17:33 |
| noonedeadpunk | you will need to set `openstack_distrib_code_name: Caracal` then | 17:33 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible-openstack_hosts/src/branch/master/defaults/main.yml#L22 | 17:33 |
| ThiagoCMC | Nice, thanks! | 17:33 |
| noonedeadpunk | I'm not really sure that caracal is out for UCA.... | 17:36 |
| noonedeadpunk | huh, so how I can possible render that readme file without gitea or anything like that... | 17:41 |
| jrosser_ | lxc is broken on 24.04 | 17:41 |
| jrosser_ | anyone who fanices fixing that should dive in | 17:41 |
| jrosser_ | ThiagoCMC: i did not see a patch from you to move OSA to reef yet? | 17:42 |
| jrosser_ | we definatly need to move to reef before making the caracal release | 17:43 |
| ThiagoCMC | Well, I've never contributed to OSA before! I'm not familiar with the procedure. I did posted the manual steps here: https://paste.opendev.org/raw/bJXKfhUBNstPINNL4xo9/ | 17:43 |
| ThiagoCMC | As you folks said... 24.04 is a distraction... lol | 17:44 |
| jrosser_ | here is the contributors quickstart https://docs.openstack.org/contributors/code-and-documentation/quick-start.html | 17:44 |
| jrosser_ | ThiagoCMC: reef needs doing | 17:44 |
| ThiagoCMC | Great! Thanks for sharing it. I'll look into it. | 17:44 |
| jrosser_ | and the rest of your paste is about making the apt pins suitable to prefer UCA, which is a different and seperable problem | 17:44 |
| ThiagoCMC | Okdok | 17:44 |
| jrosser_ | rather than just delete code we should have a proper mechanism to prefer whichever repo the deployer wants | 17:45 |
| ThiagoCMC | Yep, sounds fair... I did it like that (quick and dirty) just to test | 17:46 |
| jrosser_ | i think that switching to reef is an ideal starter contribution, as part of it is as simple as changing the setting of `ceph_stable_release` | 17:47 |
| ThiagoCMC | Ok, I'll do it! | 17:47 |
| noonedeadpunk | seems I need to also have another round on https://review.opendev.org/c/openstack/openstack-ansible-galera_server/+/914691/7 | 17:47 |
| jrosser_ | this looks pretty wrong too https://github.com/openstack/openstack-ansible-ceph_client/blob/2bb6cdf4811223406a95486be7d588d168f92c73/defaults/main.yml#L42 | 17:49 |
| noonedeadpunk | yeah | 17:50 |
| jrosser_ | i wonder if that really does install pacific client in like glance containers for AIO | 17:50 |
| jrosser_ | that would be interesting to see | 17:50 |
| noonedeadpunk | it really should not | 17:50 |
| jrosser_ | i don;t know what else would be responsible for setting up the apt source though | 17:51 |
| noonedeadpunk | https://opendev.org/openstack/openstack-ansible/src/branch/master/inventory/group_vars/all/ceph.yml#L31 | 17:51 |
| jrosser_ | aaaaahhhhhhhh | 17:51 |
| jrosser_ | ok sure | 17:51 |
| jrosser_ | noonedeadpunk: oh before i forget i think that this might be broken https://github.com/openstack/openstack-ansible/blob/master/tests/roles/bootstrap-host/templates/user_variables_magnum.yml.j2#L21 | 17:55 |
| jrosser_ | does openstack_resources handle compressed images? | 17:55 |
| jrosser_ | i think right now it uploads the xz as the image contents | 17:56 |
| noonedeadpunk | good question. somehow I thought that glance was handling that | 17:56 |
| noonedeadpunk | but checking on what we do (we internally have slightly more opinionated version of the role that needs to be refactored), and it does handle unpack.... | 17:57 |
| noonedeadpunk | tricky part of unpack is SHA verification | 17:58 |
| noonedeadpunk | jrosser_: looking through merged code - I don't see too much references merged? https://codesearch.openstack.org/?q=mcapi_vexxhost&i=nope&literal=nope&files=&excludeFiles=&repos= | 18:03 |
| noonedeadpunk | probably except ones we've sent out in ML or irc | 18:04 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Add ClusterAPI documentation to renderred docs https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916653 | 18:07 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-ops master: Add ClusterAPI documentation to renderred docs https://review.opendev.org/c/openstack/openstack-ansible-ops/+/916653 | 18:08 |
| noonedeadpunk | this renderred well: https://e586be400b0defacea9f-984ca498bb42ba0203f9a104f5aecd9c.ssl.cf2.rackcdn.com/916652/1/check/openstack-tox-docs/c898c05/docs/elk_metrics.html | 18:17 |
| -opendevstatus- NOTICE: Gerrit will be offline for a short time while we rename a project repo. https://lists.opendev.org/archives/list/service-announce@lists.opendev.org/message/KP6NCOKJEYRGFD5FS26CZPVLEKFSY2ZO/ for more details | 20:00 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!