| noonedeadpunk | Yeah, I can kinda recall doing that, I think we still do have folder for venvs on the repo container. But I can't recall waht exactly didn't went well... I guess some kind of python bindings? | 06:20 |
|---|---|---|
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_octavia master: Adjust condition for availability_zone definition https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/899127 | 06:56 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-lxc_container_create master: Use FQCN for lxc_container module https://review.opendev.org/c/openstack/openstack-ansible-lxc_container_create/+/899128 | 07:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_nova master: Drop until-complete flag for db purge https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/899129 | 07:08 |
| jrosser | noonedeadpunk: are we always connecting to lxc with ansible as the root user? | 08:15 |
| jrosser | we've talked recently about non-root user for ansible->physical_host, but what about the lxc-attach from physical_host->container | 08:16 |
| noonedeadpunk | jrosser: well, we have `container_user` param that can be used somehow... | 08:17 |
| jrosser | i was looking at the connection plugin, and done some tidying up of the code | 08:17 |
| noonedeadpunk | as of today yeah, I think with most scenarios we do lxc-attach and then su as root | 08:17 |
| jrosser | we do a whole round-trip of connection plugin to find the pid to put_file/fetch_file in the connection plugin | 08:18 |
| jrosser | like /proc/66798/root/root/.ansible/tmp/ansible-tmp-1698134553.4849596-354251-48702662267625/sourc | 08:19 |
| jrosser | and we could get rid of that perhaps, if we bind mount the ansible tmp dir | 08:20 |
| noonedeadpunk | we could do that I guess... | 08:23 |
| noonedeadpunk | But then we also can define different ansible tmp path to be more... available? rather being under $HOME | 08:27 |
| noonedeadpunk | Like jsut use /tmp/ansible or smth | 08:27 |
| jrosser | well i was trying to get rid of the PID lookup | 08:36 |
| jrosser | urgh looks like we have some widespread CI fail with ARA | 10:13 |
| kleini | I just noticed, that roles used in openstack-ansible-ops for Prometheus, node exporter and Grafana have been archived on GitHub. They got replaced with a collection. Can I extend fetched collections in OSA? would it make sense to upgrade to that collection in openstack-ansible-ops? | 11:03 |
| noonedeadpunk | kleini: sure you can do that. for this specific reasons we have user-collection-requirements and user-role-requirements | 11:31 |
| noonedeadpunk | kleini: https://docs.openstack.org/openstack-ansible/latest/reference/configuration/extending-osa.html#adding-new-collections-in-your-openstack-ansible-installation | 11:32 |
| noonedeadpunk | and yes, you can propose patches to ops repo as well. Though I gues we would need to take some care about CI in the ops... | 11:33 |
| kleini | okay, thanks for your answers. I try to stick to the old roles first and try to get some deployment up and running. currently struggling with Grafana deployment and missing data_sources | 11:35 |
| kleini | https://github.com/openstack/openstack-ansible-ops/blob/master/grafana/installGrafana.yml#L97 <- python-pymysql is from times, where OSA used Python 2, right? So nobody used Grafana deployment after OSA switched to Python 3? | 12:24 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Calculate if target is a container only once https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/899162 | 12:28 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Remove nspawn container support https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/899163 | 12:28 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Cosmetic tidy up of pid lookup function https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/899164 | 12:28 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Remove extra container check https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/899165 | 12:28 |
| jrosser | we've not used those playbooks for grafana, but we use the same role which is now archived | 12:33 |
| NeilHanlon | mornin' folks | 12:40 |
| noonedeadpunk | jrosser: is it somehow not enough to set `self.is_container = self._container_check()` only in __init__? | 12:41 |
| jrosser | could be | 12:41 |
| jrosser | though things which can affect the outcome are perhaps adjusted in set_options | 12:42 |
| jrosser | i'm not totally sure how that is supposed to be | 12:42 |
| noonedeadpunk | ah, we don't have enough context in init | 12:44 |
| noonedeadpunk | yeah | 12:44 |
| noonedeadpunk | but why then we need it in __init__... | 12:44 |
| jrosser | we can iterate these a bit if needed | 12:50 |
| jrosser | it is slightly unhelpful as __init__ also seems to retrieve those values | 12:50 |
| noonedeadpunk | but without options set it kinda makes little sense I guess | 12:51 |
| jrosser | do you think that the handling of container_name/physical_host in __init__ is bogus anyway? | 12:54 |
| jrosser | it's maybe similar to these https://github.com/ansible/ansible/blob/devel/lib/ansible/plugins/connection/ssh.py#L587 | 12:57 |
| noonedeadpunk | hm, yeah, might be... | 13:00 |
| * NeilHanlon facepalms | 13:09 | |
| NeilHanlon | I--once again--failed to convert timezones appropriately and showed up an hour early for our meeting | 13:09 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Retrieve container name and physical host via get_options https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/899166 | 13:14 |
| jrosser | well lets see what that does, it's an incremental change on top of the others to see if the settings in __init__ matter | 13:15 |
| NeilHanlon | jrosser: small bit -- s/Nono/None/g | 13:17 |
| jrosser | argh :) | 13:18 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-plugins master: Retrieve container name and physical host via get_options https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/899166 | 13:18 |
| jrosser | ty! | 13:18 |
| NeilHanlon | np | 13:18 |
| NeilHanlon | now I am wondering if `class Nono(None):` would work.. lol | 13:19 |
| jrosser | Nono sounds like a really firm version of None | 13:19 |
| jrosser | absolutely None | 13:19 |
| NeilHanlon | haha exactly | 13:19 |
| NeilHanlon | i thought of this car insurance commercial from a decade ago https://youtu.be/RtXtOuxBuvQ | 13:20 |
| noonedeadpunk | dmsimard[m]: hey! are you around? we have some weird things happening with ara1.7.0, despite your testing path passed nicely.... | 13:31 |
| jrosser | here https://zuul.opendev.org/t/openstack/build/80df1a51cee849e3961251ba50bd537a/log/job-output.txt#4064-4072 | 13:32 |
| jrosser | what is odd is this started to break just now | 13:32 |
| jrosser | but it seems that yaml lib has been at 0.17.x for some time now | 13:32 |
| noonedeadpunk | yeah... maybe some dependencies got updated.... | 13:33 |
| dmsimard[m] | noonedeadpunk: what's up ? | 13:51 |
| noonedeadpunk | smth like that https://paste.openstack.org/show/b441YNDGVTDSizdpec90/ | 13:52 |
| dmsimard[m] | Interesting, haven't seen that until now | 13:53 |
| noonedeadpunk | yep.. we neither | 13:57 |
| noonedeadpunk | start breaking just tomorrow | 13:57 |
| noonedeadpunk | *today lol | 13:57 |
| noonedeadpunk | maybe yesterday evening | 13:57 |
| dmsimard[m] | new version of ruamel.yaml I guess? | 13:59 |
| dmsimard[m] | yeah: https://pypi.org/project/ruamel.yaml/#history | 13:59 |
| dmsimard[m] | could I ask you nicely to create an issue about it ? In the meantime it seems like pinning ruamel.yaml to <0.18 could be a workaround | 14:00 |
| NeilHanlon | https://sourceforge.net/p/ruamel-yaml/code/ci/3469de4cae2c770be4ef2f1e9124bbbbb79dc7cd/ | 14:00 |
| NeilHanlon | is ruamel in u-c ? | 14:00 |
| noonedeadpunk | dmsimard[m]: sure, thanks for the quick look! | 14:01 |
| dmsimard[m] | FWIW ara doesn't /require/ ruamel, but tries to be compatible with both ruamel and python stdlib yaml: https://github.com/ansible-community/ara/blob/dde76744596b031f8a52297a71fee59ba9aeb605/ara/server/settings.py#L12-L17 | 14:01 |
| dmsimard[m] | it's a transitive dependency from dynaconf which handles the django settings | 14:01 |
| noonedeadpunk | I think we do install ruamel for other things as well.... | 14:01 |
| noonedeadpunk | And yes it's in U-C, just for ara it seems we don't use u-c? | 14:02 |
| noonedeadpunk | anyway | 14:02 |
| noonedeadpunk | It's PTG time just in case :) | 14:02 |
| NeilHanlon | :) | 14:02 |
| noonedeadpunk | So whoever interested in participating are warmly welcome to https://www.openstack.org/ptg/rooms/grizzly | 14:03 |
| dmsimard[m] | I can't make it but you guys have fun o/ | 14:04 |
| dmsimard[m] | fwiw I will be doing openstack and ansible things once again in the near future ;) | 14:04 |
| NeilHanlon | nice! | 14:04 |
| mgariepy | i'll listen in but i have something a bit later. | 14:08 |
| jrosser | noonedeadpunk: you see on the ML RDO bobcat is released, we should check again if we carry on supporting this | 14:58 |
| noonedeadpunk | Let's continue probably | 15:00 |
| jrosser | noonedeadpunk: my laptop rebooted itself :( | 15:38 |
| noonedeadpunk | mnaser: hey! are you around ? as we have CAPI driver discussion next in the list | 16:06 |
| mnaser | noonedeadpunk: sorry I’m going to be around in an hour :/ just saw this | 16:59 |
| noonedeadpunk | we've just finished though | 17:01 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_designate master: Add quorum queues support for service https://review.opendev.org/c/openstack/openstack-ansible-os_designate/+/898818 | 17:10 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Add quorum queues support for service https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/898911 | 17:11 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Add quorum queues support for service https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/898911 | 17:11 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Explicitly add constraints for ara setup https://review.opendev.org/c/openstack/openstack-ansible/+/899196 | 17:20 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible master: Explicitly add constraints for ara setup https://review.opendev.org/c/openstack/openstack-ansible/+/899196 | 17:39 |
| dmsimard[m] | ruamel is on sourceforge, huh | 20:11 |
| noonedeadpunk | ugh | 20:25 |
| dmsimard[m] | I'm pondering on the right fix, considering ara will need to continue working with both the new and the old ruamel because distros might carry whatever version | 20:29 |
| dmsimard[m] | packaging is just great | 20:29 |
| dmsimard[m] | the new ruamel is already coming to fedora: https://koji.fedoraproject.org/koji/buildinfo?buildID=2311134 | 20:32 |
| noonedeadpunk | wow, that was fast indeed.... I wonder what are changes though between 0.18 and 0.18.2 | 21:27 |
| noonedeadpunk | dmsimard[m]: well. 0.18.2 should be fixing the issue as well | 21:29 |
| dmsimard[m] | oh? | 21:30 |
| noonedeadpunk | `calling the deprecated functions now raises an `AttributeError` with the, somewhat more informative, orginal warning message. Instead of calling `sys.exit(1)`` | 21:31 |
| noonedeadpunk | and what triggered me in log that it was WARNING but rc 1 | 21:31 |
| noonedeadpunk | which kind of explains why.... | 21:31 |
| dmsimard[m] | oh, so it crashed instead of just printing the deprecation | 21:31 |
| dmsimard[m] | or rather exit 1 | 21:31 |
| noonedeadpunk | yup | 21:32 |
| dmsimard[m] | nice | 21:32 |
| noonedeadpunk | fedora was too fast to build lol | 21:32 |
| noonedeadpunk | they should have waited for another hour | 21:32 |
| NeilHanlon | lol | 21:37 |
| NeilHanlon | dmsimard[m]: share your 'huh' about the sourceforge thing lol | 21:38 |
| dmsimard[m] | ¯\_(ツ)_/¯ | 21:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!