| 11 Jan 2025 |
 Vladimír Čunát | Switched to --cores 20, for now with 5 jobs. (nproc=80, 128G RAM, /tmp on physical disk) | 09:32:10 |
| Vladimír Čunát | And restricted it to big-parallel jobs only, though I'm not 100% sure about that. | 09:36:31 |
|  sinan changed their profile picture. | 11:36:33 |
 sterni | is it just me or are release-24.11 and trunk going through their queue unusually slow? shouldn't that be mostly downloads? | 13:35:49 |
 Lily Foster | qutebrowser seems caught in crossfire, with User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) QtWebEngine/6.8.1 Chrome/122.0.6261.171 Safari/537.36 :)
(just had to use firefox to pull up a hydra build) | 14:46:20 |
 hexa | ok | 14:50:39 |
| hexa | Lily Foster: fixed | 14:51:54 |
| Lily Foster | confirmed, ty <3 | 14:52:13 |
|  Fernando Rodrigues changed their display name from SigmaSquadron (Away until 2025-01-11) to SigmaSquadron. | 15:06:44 |
| hexa | intense-heron causing eofs on builds | 15:39:59 |
| hexa | full disk | 15:41:19 |
| Vladimír Čunát | Why should trunk be mostly downloads? | 16:01:36 |
| sterni | hm you're right we haven't had a staging merge recently | 16:02:09 |
| sterni | I assumed something like that had happened since I saw always 10k queued builds a few days ago | 16:02:29 |
| Vladimír Čunát | Half an hour for kernel, around five hours for chromium. I guess that's OK latency now. | 16:03:02 |
| Vladimír Čunát | As for release-24.11, it could be finished relatively fast now that I bumped it (after its tested job succeeded). Even if some work is relatively cheap, prioritization still matters. | 16:41:22 |
| 12 Jan 2025 |
| hexa | https://github.com/NixOS/infra/issues/534 | 00:33:22 |
| 11 Jan 2025 |
|  oak 🏳️🌈♥️ changed their profile picture. | 16:45:53 |
| 12 Jan 2025 |
| hexa | The database on haumea.nixos.org failed an automatic vacuum, because it ran out of disk space. | 00:33:47 |
| 11 Jan 2025 |
| oak 🏳️🌈♥️ changed their profile picture. | 16:46:55 |
| 12 Jan 2025 |
| hexa | I've removed some expensive snapshots and started a manual full vacuum, so we can hopefully save some space. | 00:34:21 |
|  @xv1:matrix.org joined the room. | 02:51:44 |
| @xv1:matrix.org left the room. | 03:03:56 |
| hexa | this should be the ETA 🕖️ | 03:09:33 |
| hexa | vacuum done | 05:26:50 |
| Vladimír Čunát | zrepl got into a stuck state, apparently.
From zrepl status I gather that some snapshot got removed too early for it to recover. Sending a "reset" signal doesn't seem to help.
| 08:12:42 |
| hexa | there was not enough disk space to keep them around | 14:13:31 |
| hexa | there is a partically received snapshot on the receiving end | 14:15:46 |
| hexa | https://github.com/NixOS/infra/issues/534#issuecomment-2585755670 | 14:25:49 |
| 13 Jan 2025 |
| hexa | I picked two patches that enables streaming of files from the store, instead of buffering them first | 00:29:39 |
