| @cidlik:matrix.org | Hello there! | 06:19 |
|---|---|---|
| I added the change which allows to add nodes to the zuul_unreachable group from nodeset [1]. It will allow to allocate devices which are unreachable for Zuul and use them in tests. Can someone review the change please or give some callback about this idea? | ||
| [1] https://review.opendev.org/c/zuul/zuul/+/932363 | ||
| @cidlik:matrix.org | * Hello there! | 06:22 |
| I added the change which allows to add nodes to the zuul\_unreachable group from nodeset \[1\]. It will allow to allocate devices which are unreachable for Zuul and use them in tests. I would like to test embedded devices with the approach. Can someone review the change please or give some callback about this idea? | ||
| \[1\] https://review.opendev.org/c/zuul/zuul/+/932363 | ||
| -@gerrit:opendev.org- Felix Edel proposed: [zuul/zuul] 932606: Reorder filters on status page https://review.opendev.org/c/zuul/zuul/+/932606 | 12:25 | |
| @newbie23:matrix.org | Are jobs in an 'independent' queue prioritized when it comes to get a semaphore (compared to the 'dependent' queue jobs)? | 14:27 |
| This is the behaviour I seem to observe on my system, but I might be completely off mark. | ||
| @clarkb:matrix.org | I don't think so. I think the schedulers process pipelines in the order they are in your config. So you would prioritize that way? | 15:02 |
| @clarkb:matrix.org | If check comes before gate then those jobs will be given opportunity to grab the semaphores first | 15:03 |
| @clarkb:matrix.org | But also multiple schedulers makes this less deterministic | 15:03 |
| @vonschultz:matrix.org | I'm considering switching the Gerrit submit type from "cherry pick" to "rebase always". Should I change the Zuul `merge-mode` from `cherry-pick` to `rebase`? At first I thought so, but now I see that the Zuul documentation says that `rebase` "is only supported by GitHub" <https://zuul-ci.org/docs/zuul/latest/config/project.html#attr-project.merge-mode>. Should I remain with `cherry-pick` as the Zuul `merge-mode` even if I switch Gerrit to "rebase always"? | 15:04 |
| @clarkb:matrix.org | I thought we tried doing rebase on a repo or two like a decade ago and the problem was the way that Gerrit rewrites the changes makes it so that zuul can't always detect them as having "merged"? From memory I'm guess that is why zuul doesn't support that method with Gerrit. I'm not sure if using cherry pick helps | 15:37 |
| @clarkb:matrix.org | Zuulians: I'm told that workday's Zuul presentation at Open Infra Days North America was well received this week and I've been asked to encourage people to submit to the SCaLE Open Infra Days CFP for next year: https://www.socallinuxexpo.org/scale/22x/events/open-infra-days this CFP is open until November 1 | 15:39 |
| @jim:acmegating.com | jangutter: congrats! 🍻 | 15:49 |
| @flaper87:matrix.org | Is there a way to tell Zuul / Ansible to use a python from a venv instead of the global one? | 17:19 |
| We have a role with a custom library (python script) and it's running with the global interpreter. I'm wondering if there's a better way to tell it to use a venv that doesn't require changing the shebang (don't even know if that would work, TBH) | ||
| @flaper87:matrix.org | * Is there a way to tell Zuul / Ansible to use a python from a venv instead of the global one in the remote host? | 17:20 |
| We have a role with a custom library (python script) and it's running with the global interpreter. I'm wondering if there's a better way to tell it to use a venv that doesn't require changing the shebang (don't even know if that would work, TBH) | ||
| @clarkb:matrix.org | flaper87: I think nodepool images can set the python to use and that is passed to zuul and into ansible. | 17:21 |
| @clarkb:matrix.org | but that is for the remote side | 17:21 |
| @clarkb:matrix.org | on the zuul side you install ansible into venvs as part of the zuul installation (as there is one venv per ansible version installation). You can tell the zuul ansible install process to install extra libs as needed | 17:21 |
| @flaper87:matrix.org | > <@clarkb:matrix.org> flaper87: I think nodepool images can set the python to use and that is passed to zuul and into ansible. | 17:23 |
| Ah, it's in nodepool. Thanks, I just found this: https://zuul-ci.org/docs/nodepool/latest/configuration.html#attr-diskimages.python-path | ||
| @flaper87:matrix.org | > <@clarkb:matrix.org> on the zuul side you install ansible into venvs as part of the zuul installation (as there is one venv per ansible version installation). You can tell the zuul ansible install process to install extra libs as needed | 17:24 |
| We do this for Zuul and Ansible but this module/library is executed in the remote host and it's relying on the global interpreter for some reason. I may be missing something but the venvs you are referring to exist only in the executor, right? | ||
| @clarkb:matrix.org | yes the venvs I'm talking about are on the executor. I just wanted to call out there are two places where you have to consider. One is the remote test node side and one is the executor. Depending on what you are doing with custom ansible modules you may need to deal with one side or another or both | 17:25 |
| @flaper87:matrix.org | Thanks, Clark | 17:29 |
| @flaper87:matrix.org | Here's one of the exceptions I'm seeing, wondering if this is a known one. Happy to file a bug if not: | 17:42 |
| ``` | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b'Callback Exception:' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' File "/usr/local/lib/zuul/ansible/8/lib/python3.11/site-packages/ansible/executor/task_queue_manager.py", line 468, in send_callback' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' method(*new_args, **kwargs)' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' File "/var/lib/zuul/ansible/8/zuul/ansible/callback/zuul_stream.py", line 733, in v2_runner_item_on_ok' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' if isinstance(result_dict[loop_var], str):' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' ~~~~~~~~~~~^^^^^^^^^^' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b'changed: [ci-base -> localhost] => (item=None) => {' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' "censored": "the output has been hidden due to the fact that \'no_log: true\' was specified for this result",' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b' "changed": true' | ||
| 2024-10-17 17:36:06,274 DEBUG zuul.AnsibleJob.output: [e: a8016ca1f5104053901333b4900a3bb2] [build: e88b0bb213e145118098e80e3d752337] Ansible output: b'}' | ||
| ``` | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 932647: Update include-vars to use the ref checkout on ref jobs https://review.opendev.org/c/zuul/zuul/+/932647 | 18:36 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: | 20:53 | |
| - [zuul/zuul] 932449: Launcher: raise_for_status on download https://review.opendev.org/c/zuul/zuul/+/932449 | ||
| - [zuul/zuul] 932456: Add test for instance creation failure https://review.opendev.org/c/zuul/zuul/+/932456 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed on behalf of Simon Westphahl: | 20:53 | |
| - [zuul/zuul] 929273: Add support for 'min-ready' provider nodes https://review.opendev.org/c/zuul/zuul/+/929273 | ||
| - [zuul/zuul] 931779: Implement 'max-ready-age' handling https://review.opendev.org/c/zuul/zuul/+/931779 | ||
| - [zuul/zuul] 931780: Implement re-use of ready nodes w/o request https://review.opendev.org/c/zuul/zuul/+/931780 | ||
| - [zuul/zuul] 932170: Store state time for provider nodes https://review.opendev.org/c/zuul/zuul/+/932170 | ||
| - [zuul/zuul] 932179: Only expire nodes when no pending layout update https://review.opendev.org/c/zuul/zuul/+/932179 | ||
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 932659: Add most remaining openstack config options https://review.opendev.org/c/zuul/zuul/+/932659 | 22:26 | |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 932664: Restore openstack api call timer https://review.opendev.org/c/zuul/zuul/+/932664 | 23:51 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!