| 5 Apr 2026 |
|  sky1e joined the room. | 14:27:52 |
| 6 Apr 2026 |
 sterni | can someone tell me how many cores/threads elated-minsky would report? I'm currently trying to chase down a parallelism issue i have trouble I am having trouble to reproduce myself | 11:19:28 |
 hexa (signing key rotation when) | Jobs on there get 2C | 11:36:00 |
| hexa (signing key rotation when) | The whole system is 48C/96T | 11:36:56 |
| sterni | thx | 13:23:48 |
|  Hythera joined the room. | 21:49:41 |
| 8 Apr 2026 |
 ElvishJerricco | hexa (signing key rotation when): I've pushed a new version to https://github.com/NixOS/nixpkgs/pull/435781, which doesn't include the for-WIP-only jobset file used by https://hydra.nixos.org/jobset/nixos/pr-435781-systemd-stage1. So we should disable the jobset. I don't think we need another run of it. | 00:38:36 |
| ElvishJerricco | vcunat: ^ | 00:38:41 |
| hexa (signing key rotation when) | disabled | 00:39:17 |
|  tort55 joined the room. | 02:36:32 |
| 9 Apr 2026 |
 adamcstephens | I'm trying to get the revision that hydra is actively building, e.g. for nixos-unstable right now https://hydra.nixos.org/eval/1824324 . I can get this from the jobset evals API, but this is an expensive query that gathers all the build IDs as well, and can take over a minute to return a bunch of data that I don't want. Is there a more lightweight way to retrieve this info? | 13:47:27 |
| 10 Apr 2026 |
|  tbutter joined the room. | 20:07:13 |
| 11 Apr 2026 |
| hexa (signing key rotation when) | we run out of disk space on mimas | 01:20:51 |
| hexa (signing key rotation when) | gc is running but it'll take a while to meainingfully free up space | 01:21:37 |
| hexa (signing key rotation when) | NAME USED AVAIL REFER MOUNTPOINT
zroot 1.66T 21.7G 96K none
| 01:22:23 |
| hexa (signing key rotation when) | I'm going to sleep, I have a script running that restarts things once we have 200G avail disk space | 01:22:44 |
| hexa (signing key rotation when) | 💤 | 01:23:15 |
 vcunat | It works now but not much free space. I wonder why there was such a sharp spike in usage. It looks like hundreds of gigabytes within a couple hours. | 04:46:57 |
| vcunat | Ah no. The queue runner wasn't started, so we were building nothing. Starting now. | 04:52:39 |
| vcunat | * Ah no. The queue runner wasn't started, so we were building nothing. Starting now. (also evaluator, but that one wasn't so important) | 04:53:25 |
 Arian | @winter:catgirl.cloud: @winterqt:nixos.dev: i have a time-pressing issue. We enrolled the apple Macs with the MDM Push certificate you created and uploaded. We must renew them within the next 4 days. If we fail to do so the Macs turn into bricks unless we manually unenroll them now.
Could you please today log into Mosyle and follow the instructions with renewing? Or can you share in the infra vault the credentials for winter@nixos.org managed apple id?
| 08:10:27 |
| Arian | Apparently the MDM Push cert is hard-bound to the appleid they were created with. So only you can renew them ._. | 08:10:54 |
| Arian | Well they won't brick I think but we'd have to get people to manually re-enroll them most likely. But id prefer not to have to do that | 08:16:34 |
| Arian | I can also reset the password and 2fa in apple Business Manager myself as another option | 08:18:13 |
| hexa (signing key rotation when) | we're also keeping around 620G as nar-cache | 11:54:50 |
| hexa (signing key rotation when) | not sure how reasonable that is | 11:54:58 |
| vcunat | That doesn't sound reasonable to me. | 11:59:51 |
| hexa (signing key rotation when) | and 120 GB build logs | 12:00:31 |
| vcunat | Build logs get dumped to S3, so I don't expect we need a big cache here. | 12:06:26 |
| hexa (signing key rotation when) | we keep 3 years currently | 12:07:41 |
