| 23 Sep 2021 |
 Rick (Mindavi) | In reply to @mbprtpmnr:matrix.org
again, oom-kill ... Patchelf works, but the system seems too complicated for the memory | 19:19:24 |
 mbprtpmnr | Yeah, well, on my machine I have only 2GiB of RAM and is evaluated and built... | 19:21:22 |
 Mic92 | In reply to @mbprtpmnr:matrix.org
Yeah, well, on my machine I have only 2GiB of RAM and is evaluated and built... swap? | 19:21:45 |
| mbprtpmnr | This VPS hase 1GiB of RAM... Why not? | 19:21:47 |
| mbprtpmnr | In reply to @mic92:nixos.dev
swap? zramd 50%... | 19:22:11 |
| Mic92 | mbprtpmnr: i feel sorry for you. | 19:26:08 |
| mbprtpmnr | Redacted or Malformed Event | 19:28:02 |
 cransom | 1gig of ram, using just nixos-rebuild, is going to be tight, letalone evaluations/builds. | 20:01:11 |
| 24 Sep 2021 |
|  Rev. CornWallace III (novus ordo seclorum) changed their display name from Rev. CornWallace III (sun/tzu) to Rev. CornWallace III (novus ordo seclorum). | 01:01:14 |
| mbprtpmnr | In reply to @janne.hess:helsinki-systems.de
you can do multiple parallel evals on the host, but it's always limited to one host Please, can you explain how can I enable parallel evals...? | 15:49:39 |
| mbprtpmnr | * Please, can you explain how can I enable parallel evals...or how can I do that? | 15:52:22 |
| cransom | afaict it does 4 parallel by default, the config option is max_concurrent_evals. https://github.com/NixOS/hydra/blob/2bb1ba22d3717f38a162e632471407f0c09923b0/src/hydra-evaluator/hydra-evaluator.cc#L111 | 15:57:27 |
 @grahamc:nixos.org | what would y'all think about making the /build/<number> API return information about aggregate job constituents if it is an aggregate job? | 18:28:23 |
 tomberek | from Office Hours: usage of NIX_CONFIG="store = /some/path" was a configuration someone wanted to use in their services.hydra.extraEnv | 18:31:17 |
| @grahamc:nixos.org | https://github.com/NixOS/hydra/pull/1026/files | 21:08:51 |
 lukegb (he/him) | In reply to @grahamc:nixos.org
what would y'all think about making the /build/<number> API return information about aggregate job constituents if it is an aggregate job? yes please I want this | 22:53:30 |
 das_j | In reply to @grahamc:nixos.org
what would y'all think about making the /build/<number> API return information about aggregate job constituents if it is an aggregate job? Да | 23:04:20 |
| das_j | * Да пожалуйста | 23:04:53 |
| @grahamc:nixos.org | lukegb (he/him): it looks tricky so I did 1026 instead | 23:40:02 |
| lukegb (he/him) | 1026 is also cool | 23:40:19 |
| @grahamc:nixos.org | Sufficient? | 23:40:46 |
| 25 Sep 2021 |
| @grahamc:nixos.org | Also, make sure to reply on the PR if you like it / don't like it / tried it :). | 00:03:03 |
| 27 Sep 2021 |
| @grahamc:nixos.org | just re-upping, das_j and lukegb (he/him) does https://github.com/NixOS/hydra/pull/1026 implement what you need? | 14:40:26 |
| lukegb (he/him) | hey, sorry, I hadn't got around to taking a look yet | 14:41:17 |
| @grahamc:nixos.org | no worries | 14:41:37 |
| 28 Sep 2021 |
| mbprtpmnr | Another question, is there a way to have something like, one hydra master in DE and two hydra slaves in US and to have some loadbalancer in between? | 08:37:21 |
| das_j | grahamc (he/him): Maybe my knowledge of the framework is lacking here, but does $self->status_ok ever return a non-200 code? | 10:06:00 |
| das_j | I'm asking because you specify 404 as a possible response code | 10:06:30 |
| @grahamc:nixos.org | This sub never gets called if the build doesn’t exist, see the Chained thing? It calls that named sub first, which sometimes will issue a 404 | 11:25:16 |
| das_j | Oops, right. Thanks for that clarification | 11:48:27 |
