| noonedeadpunk | mornings | 07:48 |
|---|---|---|
| jrosser_ | o/ hello | 07:57 |
| admin1 | \o | 09:29 |
| opendevreview | Jonathan Rosser proposed openstack/ansible-role-python_venv_build master: Include REQUESTS_CA_BUNDLE in pip task environment https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/917838 | 09:46 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Update global pins for 2024.1 https://review.opendev.org/c/openstack/openstack-ansible/+/916792 | 09:47 |
| noonedeadpunk | jrosser_: I wonder if same should be done for venv_pip_build_env as well? I guess depends a bit on how isolated env is and if cert for internal pip is trusted... | 09:50 |
| noonedeadpunk | but need of that is very confusing... | 09:51 |
| jrosser_ | possibly - i am having some difficulty to make this testable locally | 09:51 |
| noonedeadpunk | aha | 09:51 |
| jrosser_ | i think its made more difficult becasue you get one overall shell script with gate-check-commit with a session that lives for the lifetime of that script | 09:58 |
| jrosser_ | session/shell environment | 09:58 |
| jrosser_ | so as soon as you have it fail and go poking around and re-run it, it's not the same context as you had in the first invocation of the script | 09:59 |
| noonedeadpunk | what if follow just aio process? would it be same? | 10:04 |
| noonedeadpunk | meaning bootstrap-ansible.sh, bootstrap-aio.sh, openstack-ansible setup-everything? | 10:04 |
| noonedeadpunk | as then potentially we can just insert some environment reload or smth like that to gate-check-commit | 10:05 |
| noonedeadpunk | as we actually do that already: https://opendev.org/openstack/openstack-ansible/src/branch/master/scripts/gate-check-commit.sh#L205-L208 | 10:06 |
| noonedeadpunk | right after setup-hosts.... | 10:06 |
| noonedeadpunk | and yeah... looking at gate-check-commit, this issue should not happen... | 10:11 |
| jrosser_ | looks like some tls jobs are starting to pass on 916792 now | 11:30 |
| jrosser_ | not liking that its a guess at a fix rather than actual understanding though | 11:31 |
| noonedeadpunk | yup, it's exteremely confusing | 11:34 |
| noonedeadpunk | as I actually was able to reproduce after gate-check-commit by running just the playbook (though in the same session) | 11:35 |
| noonedeadpunk | but my expectation was, that ansible should connect to host with a new connection after play end regardless | 11:35 |
| jrosser_ | yeah i dont particularly understand the lifetime of the controlpath socket | 11:36 |
| jrosser_ | if that is unique per ansible-playbook invocation | 11:37 |
| noonedeadpunk | and again - I'd expect reset_connection to actually reset it... | 11:47 |
| noonedeadpunk | will try to spawn couple of aios today as well... | 11:48 |
| nixbuilder | Question... you guys have created a couple of patches in the last week or so... 917787 and 917192. How will I know when those patches get 'published' to the main line of code... when I download the code using git? | 13:28 |
| nixbuilder | I know I did not say that very well... but I think you know what I mean. | 13:28 |
| jrosser_ | nixbuilder: so you can look at the patches in gerrit, like this https://review.opendev.org/c/openstack/openstack-ansible-os_glance/+/917787 | 13:42 |
| jrosser_ | the only way they can merge is if the CI tests pass and human reviewers approve the changes, then the CI passes for a second time | 13:43 |
| jrosser_ | that will merge the change to the master branch | 13:43 |
| jrosser_ | for a bugfix the patch will then (possibly) get cherry picked onto the stable release branches | 13:44 |
| nixbuilder | jrosser_: Thanks for the info! | 13:44 |
| jrosser_ | then the process of CI tests, review and 2nd CI test to merge has to happen | 13:44 |
| jrosser_ | the change would then merge onto the stable/<release> branch in that repo | 13:44 |
| jrosser_ | in order for that to then be "automatically" deployed in a release of openstack-ansible, it would need a further patch to be merged to update this https://github.com/openstack/openstack-ansible/blob/stable/2023.2/ansible-role-requirements.yml#L122-L127 | 13:46 |
| jrosser_ | the reason i show all this, is that it can be a long process to get that all complete (for example we have broken OVS/OVN on Rocky Linux becasue of stupid external reasons) | 13:47 |
| jrosser_ | and you have many places in openstack-ansible where there is a facility to override the git SHA, the git repo locations etc so that you can take a patch and apply it to your cloud before it is fully merged and released by openstack-ansible | 13:48 |
| jrosser_ | did we get rocky working again, before i +2 anything that might be broken? | 14:45 |
| jrosser_ | ok looks like it is ok, so we are blocked by the pip tls trouble now | 14:53 |
| noonedeadpunk | yep, rocky is fine | 14:59 |
| noonedeadpunk | I'm thinking that it might be reasonable to merge https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/917838 jsut to unblock gates and try to investigate further | 15:01 |
| noonedeadpunk | wdyt? | 15:01 |
| opendevreview | Merged openstack/openstack-ansible master: [doc] Update maintenance state for unmaintained releases https://review.opendev.org/c/openstack/openstack-ansible/+/917016 | 15:04 |
| opendevreview | Merged openstack/openstack-ansible master: [doc] Update upgrades statement for compatability matrix https://review.opendev.org/c/openstack/openstack-ansible/+/917297 | 15:10 |
| jrosser_ | noonedeadpunk: i did talk with andrewbonney about it and we both agree that this is likley caused by not starting a new shell session correctly to bring new /etc/environment content in | 15:20 |
| jrosser_ | but it is completely unobvious why the meta task doesnt work | 15:20 |
| jrosser_ | possibly the only side effect with 917838 is now giving a value to a default which was previously empty - we can add some more complexity to avoid that, but i do wonder if that is worth it | 15:22 |
| noonedeadpunk | restart sshd in gate_check_commit? :D | 15:23 |
| jrosser_ | is anyone else around to review https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/917838 | 16:26 |
| noonedeadpunk | mgariepy: NeilHanlon ? | 16:29 |
| NeilHanlon | i can review | 16:47 |
| opendevreview | Merged openstack/openstack-ansible-openstack_hosts master: Ensure codename is lowercased for cloudsig install https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/917184 | 16:51 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts stable/2023.2: Ensure codename is lowercased for cloudsig install https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/917736 | 16:55 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible-openstack_hosts stable/2023.1: Ensure codename is lowercased for cloudsig install https://review.opendev.org/c/openstack/openstack-ansible-openstack_hosts/+/917737 | 16:55 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Increase number of threads to 2 for glance in AIO https://review.opendev.org/c/openstack/openstack-ansible/+/916810 | 17:03 |
| nixbuilder | Am I correct that telemetry modules (ceilometer and gnocchi) are no longer supported in BearCat and beyond? | 18:49 |
| opendevreview | Merged openstack/ansible-role-python_venv_build master: Include REQUESTS_CA_BUNDLE in pip task environment https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/917838 | 19:14 |
| noonedeadpunk | nixbuilder: why so? | 19:23 |
| noonedeadpunk | they should be | 19:23 |
| nixbuilder | noonedeadpunk: I read a posting by someone saying that it is no longer supported... and I can't find the dashboard. That's why I asked. | 19:26 |
| noonedeadpunk | um dashboard? I don't think they ever were in horizon or anything | 19:26 |
| noonedeadpunk | but these are still active and maintained projects | 19:27 |
| noonedeadpunk | and were released for 2024.1 | 19:27 |
| noonedeadpunk | though Monasca is a bit different. And Panko was also deprecated | 19:27 |
| noonedeadpunk | but ceilomter/aodh/gnocchi should be fine | 19:27 |
| nixbuilder | noonedeadpunk: Ok... thanks. | 20:53 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!