| @ekapoun1:matrix.org | Hello, it seems setting "job.voting: False" does not impact if a waiting, depending job is skipped. Is this intended? Another way to phrase my question - is there a way to allow dependent jobs to fail without the waiting jobs to be skipped? | 17:03 |
|---|---|---|
| @ekapoun1:matrix.org | * Hello, it seems setting "job.voting: False" does not impact if a waiting, depending job is skipped. Is this intended? Another way to phrase my question - is there a way to allow dependent jobs to fail without the waiting jobs to be skipped, i.e. always run a job dependent on a job which may fail? | 17:05 |
| -@gerrit:opendev.org- Artem Goncharov proposed: [zuul/zuul] 858559: Add few gitea crd tests https://review.opendev.org/c/zuul/zuul/+/858559 | 17:22 | |
| @y2kenny:matrix.org | > <@ekapoun1:matrix.org> Hello, it seems setting "job.voting: False" does not impact if a waiting, depending job is skipped. Is this intended? Another way to phrase my question - is there a way to allow dependent jobs to fail without the waiting jobs to be skipped, i.e. always run a job dependent on a job which may fail? | 17:23 |
| doesn't that means the dependency is actually not needed? | ||
| @y2kenny:matrix.org | separate question for myself. I have been trying to debug why jobs are not triggering for a repo and got a bit stuck. Context: I have untrusted project N and project M both of which are mirror of some upstream open source projects. I have a separate untrusted project J that holds build jobs for project N and M and a config project P that define the triggers for project N and M. For project M, this setup work (that is, when changes are pushed to gerrit for project M, triggers defined in project P is executing the jobs defined in project J. But for some reason project N, the trigger doesn't seem to work. Sure enough, when I check the web UI for the projects, projects M shows a non-empty job graph, while project N's job graph is empty, even though the table at the top (listing Merge mode/check jobs, etc.) list the jobs... only the job graph is empty. | 18:11 |
| @jim:acmegating.com | Kenny Ho: i sent you a PM | 18:11 |
| @y2kenny:matrix.org | one significant different between project N and M is that project N upstream uses Zuul and each branch has .zuul.yaml in it (it is in fact the nodepool repo I just shorten it to N for ease of communication) | 18:15 |
| @y2kenny:matrix.org | I was hoping the fact that I use include: [] for both project N and project M would make them behave similarly but that does not appear to be it. | 18:17 |
| @y2kenny:matrix.org | and the trigger for project N used to work... I think it might have stopped working after 6.0 but my memory around that is a bit vague | 18:18 |
| @iwienand:matrix.org | corvus: on https://review.opendev.org/c/opendev/system-config/+/858486/ i specified | 19:13 |
| @iwienand:matrix.org | ``` | 19:13 |
| nodeset: | ||
| - nodes: [] | ||
| @iwienand:matrix.org | * ``` | 19:14 |
| nodeset: | ||
| nodes: [] | ||
| groups: | ||
| - name: a group | ||
| nodes: a-node-that-doesn't-exist-above | ||
| ``` | ||
| @iwienand:matrix.org | and got back | 19:14 |
| @iwienand:matrix.org | ```Zuul encountered a syntax error while parsing its configuration in the | 19:15 |
| repo opendev/system-config on branch master. The error was: | ||
| 'name' | ||
| ``` | ||
| @iwienand:matrix.org | do you have an intuition to where the rest of the message went? | 19:15 |
| @jim:acmegating.com | ianw: not likely recoverable; best bet is to make a unit test, comment out the configloader error trapping, and then fix it | 19:16 |
| @iwienand:matrix.org | ok, thanks. i did try a unit test and got https://paste.opendev.org/show/bFpSzw2XJWuaAN768RT0/ which seemed like about the right thing | 19:24 |
| @iwienand:matrix.org | i'll keep digging | 19:24 |
| -@gerrit:opendev.org- Clark Boylan proposed: [zuul/nodepool] 858577: Check labels use valid diskimages in config validator https://review.opendev.org/c/zuul/nodepool/+/858577 | 19:37 | |
| @chgans:matrix.org | OT: I've been "invited" by the user "test_user_nav", looks fishy to me. I'll ignore it, but wonder if it's common here on Matrix. | 21:48 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!