| 11 May 2026 |
 hexa | or rather I clicked and it keeps loading 🤔 | 12:35:56 |
 Vladimír Čunát | Most likely because eval is in late stage, so a huge postgresql transaction hogs everything involved. | 12:37:12 |
| hexa | that tracks | 12:37:46 |
 K900 | Did we merge a big rebuild into next | 13:28:07 |
| Vladimír Čunát | groff: https://github.com/NixOS/nixpkgs/pull/517946#issuecomment-4420481095 | 13:35:17 |
| Vladimír Čunát | On Hydra it looked like wasting around 7k jobs. (but part of that might've been due to other changes) Not that bad, considering that it's *-linux jobs. | 13:37:11 |
| K900 | Ouch | 13:38:14 |
| K900 | Welp | 13:38:15 |
| Vladimír Čunát | 16s to display an important man page sounded pretty bad to me. | 13:39:36 |
| Vladimír Čunát | Weird. The error did not look transient, but I can't reproduce it locally, so 🤞 | 13:46:40 |
 Bart | Not sure if this is the right channel for this, but this build failed because of disk space: https://hydra.nixos.org/build/328832145, would be nice if this would build correctly for the next unstable bump | 13:54:00 |
| Bart | * Not sure if this is the right channel for this, but this build failed because of disk space: <https://hydra.nixos.org/build/328832145>, would be nice if this would be cached for the next unstable bump | 13:54:11 |
| Vladimír Čunát | Restarted. The big NixOS channels depend on the latest kernel. | 13:59:48 |
| Vladimír Čunát | * Restarted. The big NixOS channels depend on the latest kernel IIRC. | 14:00:08 |
| hexa | wild | 14:00:08 |
| hexa | how often do we have out of space issues these days? | 14:00:21 |
| Vladimír Čunát | I do see the tmpfs space missing sometimes. | 14:00:55 |
| Vladimír Čunát | Maybe we'd better simply abandon tmpfs for the builds in the infra. | 14:01:16 |
| hexa | I thought about backing it with additional swap for spillover | 14:01:21 |
| hexa | so we'd get the best of both worlds | 14:01:32 |
| Vladimír Čunát | (we'll not in the best channel for this topic, but it doesn't really matter) | 14:01:40 |
| Vladimír Čunát | * (we're not in the best channel for this topic, but it doesn't really matter) | 14:01:46 |
| hexa | Redacted or Malformed Event | 14:02:09 |
| hexa | Redacted or Malformed Event | 14:02:11 |
| Vladimír Čunát | Maybe, if it doesn't spill commonly. | 14:02:13 |
| Vladimír Čunát | tmpfs + disk swap is less efficient than normal FS on the same drive in case it does spill. | 14:02:48 |
| hexa | is that so? | 14:03:14 |
| Vladimír Čunát | At least I think it was that way. Probably due to better work with consecutive storage of files. | 14:03:49 |
| Vladimír Čunát | Well, just conclusions here. I think it's unlikely to make i big difference for our use case. Killing tmpfs is just easier. I had done it on some of our machines already. | 14:05:14 |
 teutat3s | https://github.com/rust-lang/rust/issues/153879 | 14:25:51 |
