| 2 Apr 2025 |
 magic_rb | In reply to @phaer:matrix.org
A matrix room? You can create that yourself, then make @admin:nixos.org an admin and ask in #matrix-suggestions:nixos.org for an :nixos.org alias Thanks :) | 13:57:31 |
 emily | Gaétan Lepage:
[emily@build01:~/nixpkgs]$ uptime
20:35:14 up 3 days 20:58, 4 users, load average: 29.53, 31.60, 35.04
[emily@build01:~/nixpkgs]$ nproc
24
[emily@build01:~/nixpkgs]$ ps aux | grep nix-store
glepage 2015522 0.0 0.0 80472 5684 ? Ss 20:02 0:00 fish -c nix-store --serve --write
glepage 2015527 0.1 0.0 530700 24596 ? Sl 20:02 0:02 nix-store --serve --write
emily 2124287 0.0 0.0 6812 2188 pts/17 S+ 20:35 0:00 grep nix-store
| 20:35:38 |
 Gaétan Lepage | I have no jobs running right now | 20:50:47 |
| Gaétan Lepage | (I am connected by ssh, but am not building anything) | 20:51:24 |
| Gaétan Lepage | I have been very cautious recently with my use of the builders and noticed them being severly overloaded. | 20:52:21 |
 Grimmauld (any/all) | linux or darwin? | 20:52:23 |
| Grimmauld (any/all) | if this is linux then its me | 20:52:39 |
| Gaétan Lepage | Linux & ARM right now | 20:52:44 |
| emily | ah, it is from non-remote-builder jobs after taking a closer look. sorry about that | 20:53:00 |
| Grimmauld (any/all) | my linux x86_64 jas 4 builds left running, i'll slow on arm | 20:53:10 |
| emily | looks like both x86 and AArch64 Linux are overloaded right now | 20:53:19 |
| Grimmauld (any/all) | currently going through this list all building manually: https://github.com/NixOS/nixpkgs/pull/395499
i'll slow down | 20:53:42 |
| emily | wish we had the Nix load-limit PR, could set it globally and the machines wouldn't get like this no matter how many jobs get run | 20:53:44 |
| Gaétan Lepage | That seems like a very useful feature indeed | 20:54:20 |
| emily | might want to bump the cores/jobs down a bit, looks like we have 2–3 people running builds right now :) | 20:54:24 |
 Ihar Hrachyshka | maybe it's me. I'd love to see some guidance as to how I should build stuff though. | 20:54:25 |
| Grimmauld (any/all) | the x86_64 one should really be fine though, it only has 1 job running from me at this point | 20:54:27 |
| Grimmauld (any/all) | 
Download image.png | 20:54:44 |
| emily | I'll do what I was going to do on the Darwin one since that one is idle 😅 | 20:54:49 |
| Grimmauld (any/all) | will slow on the aarch64 though | 20:54:50 |
| Grimmauld (any/all) | Sorry! | 20:55:08 |
| Gaétan Lepage | 
Download clipboard.png | 20:55:09 |
| emily | you want cores × jobs < nproc always although when multiple people are using them that pretty much has to be divided by that | 20:55:31 |
| Grimmauld (any/all) | when i started this work it was completely idle | 20:55:33 |
| emily | thankfully the AArch64 one has so many cores that you can probably just pick something conservative and still have it finish faster than anything else would :P | 20:55:45 |
| emily | no worries. it's hard to load balance | 20:55:56 |
| Ihar Hrachyshka | I do 1 job and 20 cores on aarch64 but it's pytorch... | 20:55:56 |
| emily | and sorry again to Gaétan Lepage for the dodgy grep :) | 20:56:04 |
| Gaétan Lepage | Ideally we would need a scheduler like SLURM for those shared systems | 20:56:07 |
| Grimmauld (any/all) | thats exactly why i am building all those packages on aarch64-linux, it just is the fastest builder | 20:56:13 |
