| @cidlik:matrix.org | > <@mhuin:matrix.org> In the meantime you may want to investigate the POST_FAILUREs; there was a problem in a post playbook - oftentimes it's an unset variable or expected artifacts that aren't present. | 07:06 |
|---|---|---|
| I created DNM [empty commit](https://review.opendev.org/c/zuul/zuul-jobs/+/901294) to understand, are my changes a reason of POST_FAILURE. The same jobs are failing. And due to this job i cannot read job logs. Is the way to read some logs from Zuul? In our Zuul instance all service logs place into Web on 8001 port. | ||
| @cidlik:matrix.org | > <@mhuin:matrix.org> In the meantime you may want to investigate the POST_FAILUREs; there was a problem in a post playbook - oftentimes it's an unset variable or expected artifacts that aren't present. | 07:07 |
| * I created DNM [empty commit](https://review.opendev.org/c/zuul/zuul-jobs/+/901294) to understand, are my changes a reason of POST\_FAILURE. The same jobs are failing. And due to this job i cannot read job logs. Is the way to read some logs from Zuul? E.g., in our Zuul instance all service logs place into Web on 8001 port. | ||
| @cidlik:matrix.org | * I created DNM [empty commit](https://review.opendev.org/c/zuul/zuul-jobs/+/901294) to understand, are my changes a reason of POST\_FAILURE. The same jobs are failing. And due to this error i cannot read job logs. Is the way to read some logs from Zuul? E.g., in our Zuul instance all service logs place into Web on 8001 port. | 07:20 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 901470: Add safety when setting refs https://review.opendev.org/c/zuul/zuul/+/901470 | 07:59 | |
| @fungicide:matrix.org | @cidlik:matrix.org: i don't see any post_failure results in https://zuul.opendev.org/t/zuul/buildset/aeb5e259432e4aa7b3ba10548793f4f3 so it's possible you're confusing the voting failures from the official zuul with advisory build results that are reported by third-party ci (software factory's zuul, in this case) | 13:35 |
| @fungicide:matrix.org | also note that which jobs get run for your other change will be somewhat contextual, whereas that empty change ran every possible job and some are probably bitrotten because they never get run on real changes these days | 13:37 |
| @fungicide:matrix.org | unrelated, it looks like some commit in the past few weeks has changed (broken?) the branch filter for the builds page in the dashboard when it comes to filtering master branch builds in timer-triggered pipelines in opendev's deployment | 13:59 |
| @fungicide:matrix.org | for whatever reason, master branch builds show the branch as "refs/heads/master" instead of just "master" and trying to filter on either returns an empty set. other branches don't seem to exhibit the same behavior, they don't have the "refs/heads/" prefix and can be filtered normally by name | 13:59 |
| @fungicide:matrix.org | has anyone else observed this recently? i haven't gone digging in the commit history to see if anything jumps out at me as possibly related | 14:00 |
| @fungicide:matrix.org | users report it was working as of two weeks ago, so likely changed in one of our recent weekend upgrades | 14:01 |
| @jim:acmegating.com | fungi: i see a mixture of master and refs/heads/master going back to at least early october | 14:14 |
| @jim:acmegating.com | so whatever that part is, i don't think it's new | 14:15 |
| @jim:acmegating.com | fungi: i think this may be schema-migration related; i'll look into it. | 14:43 |
| @jim:acmegating.com | fungi: okay i think i understand the issue and how to fix. first, if you look at this page: https://zuul.opendev.org/t/openstack/buildsets?project=openstack%2Fneutron&pipeline=periodic&skip=0 this is what that query returns in the db: https://paste.opendev.org/show/bOrzPYpK9LxKWUTEVsDe/ | 15:01 |
| @jim:acmegating.com | when we look for an entry in the ref table for a periodic job, the only data we query for are (project,ref). branch isn't involved in the query because it's duplicative of ref (but it is a column in the table, and that's what gets used for the web ui branch filter queries). there should be only one ref table entry for a given project-ref. but when we collapsed the duplicates during the schema migration, we happened to pick, for the master branch, a row with no branch value. probably, because as you can see from the url, it's ancient. | 15:05 |
| @jim:acmegating.com | so i think to fix, we need to go over all of the zuul_ref entries, and any where the ref column starts with 'refs/heads' and the branch column is the empty string, we should fill in the branch column with the part after refs/heads. | 15:06 |
| @jim:acmegating.com | we can do that in a migration and make it automatic for us and anyone else in this situation. i think that can happen on-line. | 15:07 |
| @fungicide:matrix.org | aha, and yeah since it's just adding data into empty fields it sounds safe as an online migration | 15:11 |
| -@gerrit:opendev.org- Zuul merged on behalf of James E. Blair https://matrix.to/#/@jim:acmegating.com: [zuul/zuul] 901510: Implement server-side filtering and pagination of config errors https://review.opendev.org/c/zuul/zuul/+/901510 | 16:05 | |
| @jim:acmegating.com | i think we're looking at something like this: `update zuul_ref set branch=substring(ref from 12) where branch='' and ref like 'refs/heads/%';` | 20:02 |
| @jim:acmegating.com | fungi: ^ | 20:02 |
| @jim:acmegating.com | on a copy of opendev's db, that results in Query OK, 30724 rows affected (3.30 sec) which seems about right. | 20:03 |
| @fungicide:matrix.org | corvus: as a manual query or is that something that should be included in zuul's db migrations? | 20:04 |
| @jim:acmegating.com | migration | 20:04 |
| @jim:acmegating.com | (but also, we could manually run it at any time since it's forward compatible) | 20:05 |
| @fungicide:matrix.org | makes sense, and the number of records matched seems within the realm of probability | 20:17 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 901624: Fix some old entries in zuul_ref table https://review.opendev.org/c/zuul/zuul/+/901624 | 20:29 | |
| @jim:acmegating.com | fungi: ^ | 20:29 |
| @fungicide:matrix.org | thanks! | 20:31 |
| -@gerrit:opendev.org- James E. Blair https://matrix.to/#/@jim:acmegating.com proposed: [zuul/zuul] 901636: Further fixes to check deduplication https://review.opendev.org/c/zuul/zuul/+/901636 | 23:06 | |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!