| 15 Jun 2022 |
 @linus:schreibt.jetzt | In reply to @kenran_:matrix.org
Hum, after having reset the server to the rescue mode and having gone back to kexec, I suddenly got a message that the server would restart in 5 minutes, which it did, killing my progress. Any idea what could have triggered that reboot? Yeah that's built into some of the kexec images... | 16:56:24 |
| @linus:schreibt.jetzt | Not sure which ones exactly but I don't think that part is in nixpkgs, so probably check whatever's not from nixpkgs | 16:57:01 |
| 16 Jun 2022 |
|  Bryan Honof changed their profile picture. | 08:43:13 |
 Dandellion | It seems to only happeb when I manually start the evaluation. When it gets ran automatically it completes and starts building | 12:33:26 |
| Dandellion | * It seems to only happen when I manually start the evaluation. When it gets ran automatically it completes and starts building | 12:33:32 |
 Robert Hensing (roberth) | is this the right place to request a hydra.nixos.org jobset change? | 17:53:45 |
| Robert Hensing (roberth) | I'd like for https://hydra.nixos.org/jobset/nixos/python-test-refactoring to be renamed to nixos-test-changes and fetch from a branch of the same name on NixOS/nixpkgs | 17:55:21 |
 hexa | Robert Hensing (roberth): no, #infra:nixos.org is the room for that 🙂 | 21:18:10 |
| 18 Jun 2022 |
|  wiredhikari joined the room. | 10:05:35 |
 sterni (he/him) | seems to me the hydra test suite is flaky, almost fails consistently in ofborg, but when executed locally it always succeeds | 19:19:02 |
| sterni (he/him) | my best guess would be that it falls over on busy machines | 19:19:18 |
 Rick (Mindavi) | In reply to @sternenseemann:systemli.org
seems to me the hydra test suite is flaky, almost fails consistently in ofborg, but when executed locally it always succeeds Yeah :( I saw that already, seems to time out regularly | 19:33:37 |
| Rick (Mindavi) | Maybe the timeouts should be longer or configurable | 19:34:14 |
| 20 Jun 2022 |
|  Florian | W3F changed their display name from Florian | OoO till 20.06. to Florian | W3F. | 08:59:21 |
|  toraritte joined the room. | 14:26:19 |
| 22 Jun 2022 |
 @trofi:matrix.org | I'd like to grep for a particular string in build logs of failed jobs for a given jobset. Is there a hydra CLI for that? Or maybe there is a syntax for web UI to do the same? An jobset would be https://hydra.nixos.org/eval/1767990?filter=x86_64-linux&compare=1767840&full=1 , an example string to search would be multiple definition of | 07:16:47 |
| @trofi:matrix.org | * I'd like to grep for a particular string in build logs of failed jobs for a given jobset. Is there a hydra CLI for that? Or maybe there is a syntax for web UI to do the same? An example jobset would be https://hydra.nixos.org/eval/1767990?filter=x86_64-linux&compare=1767840&full=1 , an example string to search would be multiple definition of | 07:17:05 |
 das_j | In reply to @trofi:matrix.org
I'd like to grep for a particular string in build logs of failed jobs for a given jobset. Is there a hydra CLI for that? Or maybe there is a syntax for web UI to do the same? An example jobset would be https://hydra.nixos.org/eval/1767990?filter=x86_64-linux&compare=1767840&full=1 , an example string to search would be multiple definition of You can only crawl Hydra and ask it for all logs, there is no API for this | 08:14:09 |
| das_j | In reply to @trofi:matrix.org
I'd like to grep for a particular string in build logs of failed jobs for a given jobset. Is there a hydra CLI for that? Or maybe there is a syntax for web UI to do the same? An example jobset would be https://hydra.nixos.org/eval/1767990?filter=x86_64-linux&compare=1767840&full=1 , an example string to search would be multiple definition of * You can only crawl Hydra and ask it for all logs, there is no API for searching logs | 12:34:56 |
 ajs124 | also, the logs are stored compressed in the filesystem, not in the database with some kind of index. so we can't just add an API endpoint that queries the database for this. | 12:55:05 |
 Sandro | Filesystem also can be S3, right? | 15:38:14 |
| das_j | wat | 15:38:37 |
| das_j | no, s3 is not a filesystem | 15:38:49 |
| das_j | If you mean store_uri - then yes, that can be a s3 url | 15:39:48 |
| @trofi:matrix.org | Unqueryable logs makes sense. I hoped for a blessed tool to fetch build logs (possibly with a bit of caching support). | 16:20:04 |
| ajs124 | you'd probably need to send an api request for each build in the eval, with that you can query the binary cache for the logs | 16:24:43 |
| Sandro | In reply to @janne.hess:helsinki-systems.de
If you mean store_uri - then yes, that can be a s3 url I meant hydra stores logs on s3 | 19:51:53 |
| 23 Jun 2022 |
| @linus:schreibt.jetzt | In reply to @trofi:matrix.org
Unqueryable logs makes sense. I hoped for a blessed tool to fetch build logs (possibly with a bit of caching support). nix log is a blessed tool for fetching build logs | 00:40:47 |
| @linus:schreibt.jetzt | It's not hydra-aware though | 00:40:56 |
| Dandellion | At least when I only had one job, now that I have two it never completes an evaluation | 04:07:00 |
