| 15 Nov 2022 |
|  vdot0x23 joined the room. | 18:52:26 |
| 16 Nov 2022 |
|  @omlet:matrix.org joined the room. | 20:35:36 |
|  @thobys:inpt.fr joined the room. | 22:34:06 |
| 17 Nov 2022 |
|  search-sense joined the room. | 09:18:07 |
| 18 Nov 2022 |
|  @rimuru:gentoo.chat changed their profile picture. | 22:07:41 |
| 25 Nov 2022 |
 terwiz | Just noticed this in Hydra manual: "At the moment, Hydra runs only on GNU/Linux (i686-linux and x86_64_linux)."
So, is it futile attempt to run a hydra slave on an arm server? | 14:44:51 |
 @linus:schreibt.jetzt | terwiz: I've been running it on my (aarch64) honeycomb lx2 for half a year now :) | 14:45:48 |
| @linus:schreibt.jetzt | since https://github.com/NixOS/hydra/pull/1262 merged it's also being built and tested on aarch64 as well | 14:46:04 |
 Rick (Mindavi) | Lol, my hack still going strong 😜 | 14:48:02 |
| terwiz | Ok, thanks | 14:48:20 |
 ajs124 | In reply to @terwiz:matrix.org
Just noticed this in Hydra manual: "At the moment, Hydra runs only on GNU/Linux (i686-linux and x86_64_linux)."
So, is it futile attempt to run a hydra slave on an arm server? slave as in build machine? those can have basically any architecture afaik | 16:12:51 |
| @linus:schreibt.jetzt | Oh yeah missed that. Indeed, if Nix runs on it you're good | 16:17:45 |
|  @tired:fairydust.space joined the room. | 22:04:40 |
| 26 Nov 2022 |
|  @/yvan:matrix.org joined the room. | 18:09:39 |
 magic_rb |
is there any way to deal with hydra eating upwards of 8GB of memory when evaluating my flake? I would like it to build all my systems but 8GB of memory is way too much for that, I think the issue Nix not really GCing that well if I try to `nix eval` all the systems at once, one by one when they're perfectly fine though… | 19:10:42 |
| @linus:schreibt.jetzt | magic_rb: add some swap. Most of that memory isn't actually being used most of the time so it can be swapped out. | 19:15:40 |
| magic_rb |
won't help in my case as cgroup still count it towards the overall usage and it'll get killed anyway, if i increase the limit for the whole container i would then run the risk of OOMing my system due to a really memory intensive build that does actually need the ram actively | 19:16:55 |
| magic_rb |
But it is true that i do not actually have any swap currently, i should probably get some but it won't help in this case | 19:18:26 |
|  @ahsmha:matrix.org changed their display name from rh to ahmed. | 19:19:32 |
| magic_rb |
I could get an SLC ssd for both L2 cache and another one for swap but wow are they expensive | 19:23:27 |
| @linus:schreibt.jetzt | In reply to @magic_rb:matrix.redalder.org
won't help in my case as cgroup still count it towards the overall usage and it'll get killed anyway, if i increase the limit for the whole container i would then run the risk of OOMing my system due to a really memory intensive build that does actually need the ram actively Limit the RAM for nix-daemon where the builds happen, but not for hydra-evaluator? | 19:24:18 |
| @linus:schreibt.jetzt | In reply to @magic_rb:matrix.redalder.org
But it is true that i do not actually have any swap currently, i should probably get some but it won't help in this case zramSwap is free :D | 19:24:40 |
| magic_rb |
I would have to move the nix-deamon out of the container into a separate container in that case | 19:24:45 |
| @linus:schreibt.jetzt | Ooh you run it all in a container? | 19:25:05 |
| magic_rb |
which may be a reasonable thing to do, but itll increase complexity a lot | 19:25:12 |
| magic_rb |
yeah i do | 19:25:14 |
| @linus:schreibt.jetzt | What about limiting nix-daemon within the container? | 19:25:30 |
| magic_rb |
nomad would still have issues with the ram amounts, i would have to keep memory down to 4GB then increase memory max above 8 to 12 which is a really bad idea because nomad schedules based on memory not memory max | 19:26:30 |
| magic_rb |
i may just have to split it up, nix-daemon, hydra-http and hydra-evaluator | 19:26:52 |
| @linus:schreibt.jetzt | Oh yeah if you're running in nomad that sounds sensibel | 19:27:08 |
