| 12 Mar 2026 |
| Theuni changed their display name from Christian Theune to Theuni. | 07:18:47 |
sterni | has anyone made a script yet that fetches the remainnig queued jobs for a Hydra jobset eval and builds them locally? | 14:14:14 |
vcunat | 🤔 I think I've seen something meant for CI which builds only the stuff which can't be substituted. | 14:19:49 |
vcunat | Found it: https://github.com/Mic92/nix-build-uncached | 14:20:25 |
vcunat | (I'm not sure what exactly you want and I haven't even used this one, though.) | 14:21:40 |
sterni | i guess that sort of works but is annoying if you have to eval the jobset locally again instead of fetching the information from Hydra | 21:14:05 |
sterni | to be fair though my use case is a bit niche, I sometimes want to build the 50 or so remaining builds from haskell-updates since you can get stuck that way for days feels like sometimes when your jobset gets crowded out | 21:14:52 |
vcunat | I see. Let me bump those now, as there are few and won't slow the other big work much (e.g. staging-next). | 21:31:03 |
vcunat | Ideally you'd somehow fetch just the names of jobs which haven't finished and then evaled (+built) just those. | 21:32:16 |
vcunat | That might be not too hard. | 21:33:03 |
sterni | yeah it's probably just a shell script with jq | 21:45:28 |
sterni | thanks :) | 21:45:39 |
| 14 Mar 2026 |
| amadaluzia -> 4d2.org changed their display name from amadaluzia to amadaluzia[uorg]. | 18:49:51 |
| amadaluzia joined the room. | 19:29:26 |
| amadaluzia -> 4d2.org changed their display name from amadaluzia[uorg] to amadaluzia -> 4d2.org. | 21:21:29 |
| 17 Mar 2026 |
Mic92 | My calendar says today is hydra meeting? das_j ? | 16:53:04 |
Mic92 | We have now a hydra meeting anyway | 17:26:27 |
| 18 Mar 2026 |
John Ericson | Simon Hauser: whose responsibility is the build-logs dir these days? | 16:54:26 |
hexa | systemd.services.hydra-prune-build-logs = {
description = "Clean up old build logs";
startAt = "weekly";
serviceConfig = {
User = "hydra-queue-runner";
Group = "hydra";
ExecStart = lib.concatStringsSep " " [
(lib.getExe pkgs.findutils)
"/var/lib/hydra/build-logs/"
"-ignore_readdir_race"
"-type"
"f"
"-mtime"
"+${toString (3 * 365)}" # days
"-delete"
];
};
};
| 21:17:19 |
hexa | imo nix itself needs to gc that | 21:17:26 |
hexa | * imo nix an hydra need a gc for that | 21:17:49 |
hexa | because nix just has it in /nix/var/log/nix instead | 21:18:10 |
hexa | * imo nix and hydra both need a gc for that | 21:18:19 |
| 19 Mar 2026 |
vcunat | We see lots of abortions from S3 uploads. Maybe this channel is also suitable, but I posted on https://matrix.to/#/!RROtHmAaQIkiJzJZZE:nixos.org/$94yUJZCsSKev_skNS2jOMGGOrrHUBch2FqYC9CAfPwY?via=nixos.org&via=matrix.org&via=nixos.dev | 08:13:14 |
| zmitchell joined the room. | 14:26:33 |
BMG | I have some time in the next couple months I can dedicate to importing state from s3 for references into hydra and start building GC. | 18:11:22 |
BMG | Need to check the latest inventory to see if that 100GB we deleted is still showing up. Been over a month now right? | 18:11:45 |
jappie | 28 days according to this post: https://discourse.nixos.org/t/garbage-collecting-cache-nixos-org/74249/10?u=jappie3 | 19:55:22 |
jappie | * 28 days according to this post: https://discourse.nixos.org/t/garbage-collecting-cache-nixos-org/74249/1 | 19:55:27 |
jappie | * 28 days according to this post: https://discourse.nixos.org/t/garbage-collecting-cache-nixos-org/74249/10 | 19:55:28 |