| 6 Mar 2025 |
 polygon_ | Hello, is there a way to get lists (newly failing jobs, still failing jobs) as JSON or another easily machine readable format? E.g. https://hydra.nixos.org/eval/1810654?full=1#tabs-now-fail | 18:09:04 |
 K900 | No | 18:10:55 |
 @janne.hess:helsinki-systems.de | In reply to @k900:0upti.me
No Actually I crawl them for zh.fail and parse the HTML. Maybe I can just serve the cache files with nginx 🤔 | 19:25:49 |
| polygon_ | Do you happen to also crawl the logs? I noticed that quite some packages failed (and less popular ones still do) after moving to GCC14 due to some warnings that got turned to errors. I compiled a list of packages that failed in the first eval after the GCC change and a current eval. Identified 400 packages that failed first then and still fail now. Would the Hydra people be unhappy if I pulled all the logs for that, the ones caused by these warnings should be easily identifiable. | 19:56:25 |
| K900 | Logs are available on the S3 | 19:56:59 |
| K900 | You can pull from there | 19:57:02 |
| K900 | That's basically free | 19:57:11 |
| polygon_ | Do I just use the Hydra links (e.g. https://hydra.nixos.org/build/291787217/nixlog/1/raw ) or something else? | 19:58:36 |
| polygon_ | Basically got a list of build IDs | 19:58:45 |
| K900 | Those Hydra links redirect to https://cache.nixos.org/log/3ipspns3k0gk8v9yp775w0blg8l6mm5w-clasp-2.6.0.drv | 19:59:07 |
| K900 | That last part is just the drv name | 19:59:20 |
| K900 | Or do you not have those? | 19:59:28 |
| polygon_ | I do not think so, is the redirect expensive on Hydra? I just have everything that is in the list of failed builds. | 19:59:53 |
| K900 | Expensive-ish | 20:00:07 |
| K900 | If you have the attrnames, just run an eval locally | 20:00:19 |
| K900 | And cross-reference | 20:00:24 |
| K900 | A full nixpkgs eval takes 10-15 minutes and is notably free for Hydra | 20:00:54 |
 ghpzin (moved to @ghpzin:envs.net) | I assumed it would use json api endpoints for that, assuming you still crawl every build by id after pulling list from eval. | 20:01:57 |
| @janne.hess:helsinki-systems.de | Let me check | 20:02:36 |
| ghpzin (moved to @ghpzin:envs.net) | * I assumed it would use json api endpoints for that, assuming you still crawl every build by id after pulling list from eval.
https://github.com/NixOS/hydra/blob/master/doc/manual/src/api.md | 20:02:47 |
| @janne.hess:helsinki-systems.de | given a build ID, I can give you lines in the form of auctex.aarch64-darwin 212687937 auctex-12.3 aarch64-darwin Failed. Would that be helpful? | 20:03:11 |
| @janne.hess:helsinki-systems.de | Redacted or Malformed Event | 20:03:22 |
| polygon_ | I would need a mapping to the store path, then I can get them from S3 directly | 20:03:42 |
| @janne.hess:helsinki-systems.de | the second field is the build id, maybe that helps? | 20:03:58 |
| ghpzin (moved to @ghpzin:envs.net) | * I assumed it would use json api endpoints for that, if you still crawl every build by id after pulling list from eval.
https://github.com/NixOS/hydra/blob/master/doc/manual/src/api.md | 20:04:40 |
| polygon_ | I already got all of that from the HTML list, but for the log I'd need the drv-path with the store hash: 3ipspns3k0gk8v9yp775w0blg8l6mm5w-clasp-2.6.0.drv | 20:05:26 |
| polygon_ | I probably go with evaluating locally, it's just a subset of 400 packages | 20:06:17 |
| 7 Mar 2025 |
| polygon_ | Thanks for the help yesterday, list with still failing builds after gcc-14 update is complete: https://polygon.github.io/fix-nixpkgs-gcc14/ | 15:43:18 |
| 10 Mar 2025 |
|  @aftix:matrix.org joined the room. | 05:18:27 |
| @aftix:matrix.org | what's the magic to get LDAP auth to work on hydra? looking at the debug logs, my hydra can bind, correctly find the LDAP user, correctly find the LDAP groups, and then send the correct password to re-bind as the LDAP user (which works with ldapwhoami), but logging into my admin account gives "bad username or password" and my user account gives "no attribute cn for User" (there is a CN attribute according to ldapsearch and my hydra configuration doesn't mention CN anyway, just name) | 14:24:37 |
