| 30 Sep 2022 |
 @linus:schreibt.jetzt | das_j: because builds has a non-nullable jobset_id column which has to refer to a jobset | 13:30:58 |
 das_j | Just create any jobset and use that when manually inserting ;) | 13:31:24 |
| @linus:schreibt.jetzt | hmmmm | 13:31:36 |
| @linus:schreibt.jetzt | it's horrible, I love it | 13:31:56 |
| das_j | You are hacking on the hydra db directly, what did you expect :D | 13:32:26 |
|  @0b11:matrix.org removed their display name 0x53. | 17:33:19 |
| @0b11:matrix.org left the room. | 17:33:27 |
| 1 Oct 2022 |
|  @mangoiv.:matrix.org joined the room. | 11:35:21 |
| 5 Oct 2022 |
|  marksisson joined the room. | 02:05:26 |
|  terwiz joined the room. | 07:11:28 |
| terwiz | Hi! What can I do when hydra refuses to re-evaluate a jobset? I mean, it says evaluation is running for few seconds and last checked time changes, but nothing else happens. There are changes in the repo. (Last build failed on a missing patch and I have included a commit that should fix that) | 07:19:52 |
| das_j | In reply to @terwiz:matrix.org
Hi! What can I do when hydra refuses to re-evaluate a jobset? I mean, it says evaluation is running for few seconds and last checked time changes, but nothing else happens. There are changes in the repo. (Last build failed on a missing patch and I have included a commit that should fix that) Did you force push? | 07:21:49 |
| das_j | You can also try clearing the scm cache in the admin menu | 07:22:19 |
| terwiz | In reply to @janne.hess:helsinki-systems.de
Did you force push? Yes | 07:22:41 |
| das_j | In reply to @terwiz:matrix.org
Yes Then drop the cache. Hydra behaves pretty strange when force pushing unfortunately | 07:23:03 |
| das_j | It expects linear history | 07:23:14 |
| das_j | And I guess it tried git log between your old and new commit and they don't have history between them as they have the same parent | 07:23:46 |
| terwiz | I do have actually another branch that is merged, but it would not re-evaluate that either. But then again, if I change the branch, I don't know what it is comparing in that case. | 07:25:54 |
| das_j | Neither do I from the top of my head :D | 07:26:35 |
| das_j | Dropping the cache should do it, if not then I can't really help you as I'm currently on the road | 07:27:05 |
| terwiz | In reply to @janne.hess:helsinki-systems.de
Then drop the cache. Hydra behaves pretty strange when force pushing unfortunately Ok, I cleared failed builds cache and VCS cache, is there yet some other cache to clear? | 07:27:47 |
| das_j | None I'm aware of | 07:30:49 |
| terwiz | In reply to @janne.hess:helsinki-systems.de
Dropping the cache should do it, if not then I can't really help you as I'm currently on the road Nope dropping caches didn't change anything. Thanks for the info about force pushing anyway. | 07:32:18 |
| terwiz | I cloned the jobset and got it building under the new name. If there is any idea how to "reset" a jobset without doing this, please let me know. | 08:05:22 |
 @v0id:nltrix.net | re-saving also triggers. | 19:06:29 |
| @v0id:nltrix.net | also check /queue-runner-status | 19:06:42 |
| @v0id:nltrix.net | systemd restart hydra-queue-runner if it stucks | 19:06:57 |
| terwiz | Well, restarting hydra-queue-runner did something, some old builds popped up running. Still had to create new jobset to get build going. | 20:28:10 |
| terwiz | Starting to think there is something wrong in the hydra config, or there are some artefacts left lingering that somehow confuse hydra. (This time I was using branch without force pushes) | 20:31:17 |
| 6 Oct 2022 |
| @v0id:nltrix.net | In reply to @terwiz:matrix.org
Starting to think there is something wrong in the hydra config, or there are some artefacts left lingering that somehow confuse hydra. (This time I was using branch without force pushes) afaik, there is debugServer = true; option on hydra service config, mite help. | 22:16:10 |
