| frickler | I'm a bit confused by the above nodepool patch, from what I see it went immediately into gate with a V-1 when tristan approved it, is that queue configured differently than others? | 07:09 |
|---|---|---|
| frickler | the usual behaviour from the recheck should've been another check run and moving into gate only when the V+1 is submitted | 07:10 |
| frickler | also could ppl please review https://review.opendev.org/c/zuul/nodepool/+/834152 in order to unblock openstacksdk 1.0? ianw in particular if you have time | 07:11 |
| fungi | frickler: nodepool is in zuul's tenant, so it uses the pipeline configuration from there | 11:40 |
| fungi | it's been configured not to require passing check results in order to enqueue in gate since it was first created | 11:41 |
| fungi | it also makes use of the "supercedes" pipeline option to cancel builds in check if there are any running when that happens | 11:44 |
| fungi | compare https://opendev.org/zuul/project-config/src/branch/master/zuul.d/pipelines.yaml to https://opendev.org/openstack/project-config/src/branch/master/zuul.d/pipelines.yaml and you'll see there's no Verified: [1, 2] for username: zuul in the approval section of the gate pipeline's gerrit trigger | 11:48 |
| fungi | you'll also see the gate pipeline has a comment-added trigger matching recheck or reverify | 11:51 |
| frickler | fungi: thx, guess that scenario just doesn't happen often enough that I'd noticed it before | 13:10 |
| fungi | yeah, the "clean check" configuration the openstack tenant uses is a bit of an anti-pattern. it's there to mitigate traffic congestion from gate resets/thrash/churn in openstack projects' long queues, but other projects usually don't have the same combination of high change volume, long-running tests and jobs which integrate dozens of different projects | 14:15 |
| frickler | another reason is to see results from non-voting jobs, if one has a sequence of patches merged without them being run, it may get difficult to find out what has been breaking them. but yes, the usefulness of this varies. good thing that zuul is flexible enough to support various options | 14:38 |
| fungi | right. if you want to see the non-voting job results in zuul's tenant, don't approve the change until you get those results | 15:45 |
| opendevreview | Steve Baker proposed openstack/diskimage-builder master: Move grub-install to the end, and skip for partition images https://review.opendev.org/c/openstack/diskimage-builder/+/826976 | 19:43 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!