!RROtHmAaQIkiJzJZZE:nixos.org

NixOS Infrastructure

479 Members
Next Infra call: 2024-07-11, 18:00 CEST (UTC+2) | Infra operational issues backlog: https://github.com/orgs/NixOS/projects/52 | See #infra-alerts:nixos.org for real time alerts from Prometheus.149 Servers

Load older messages


SenderMessageTime
20 Jun 2026
@joerg:thalheim.ioMic92https://hydra.nixos.org/build/331695810/nixlog/1 matches the description07:13:12
@joerg:thalheim.ioMic92If it's not finished in 90 min, I will deploy07:58:18
@joerg:thalheim.ioMic92I also figured out why kernels are not build on non-big-parallel machines07:59:05
@joerg:thalheim.ioMic92The json parser for derivations wasn't taking structured attrs into account07:59:37
@joerg:thalheim.ioMic92* I also figured out why kernels are build on non-big-parallel machines07:59:48
@hexa:lossy.networkhexaugh, kernel builds on 2 cores are no fun08:08:19
@joerg:thalheim.ioMic92Classic bug. I did report a similar issue a month ago to nixbuild.net08:09:40
@hexa:lossy.networkhexaRedacted or Malformed Event08:10:47
@hexa:lossy.networkhexaI don't think we want to under these circumstances08:11:08
@hexa:lossy.networkhexaif we restart it gets a big-parallel build slot and will be done much quicker08:12:01
@joerg:thalheim.ioMic92Okay, than lets roll08:17:29
@hexa:lossy.networkhexado you need any help deploying?08:17:49
@joerg:thalheim.ioMic92No. I have all closures already cached08:18:05
@hexa:lossy.networkhexacool08:18:08
@joerg:thalheim.ioMic92Fix works: https://hydra.nixos.org/machines08:40:22
@joerg:thalheim.ioMic92hopeful-rivest now gets the linux build instead08:40:34
@hexa:lossy.networkhexaalso looks like jobs stay much shorter in receiving outputs now08:51:08
@hexa:lossy.networkhexaso that works, too at first glance08:51:17
@joerg:thalheim.ioMic92Yeah. Ingestion is now a bit busier. But not as busy as in the old days.08:53:30
@hexa:lossy.networkhexaI think we need to tune the load on hopeful-rivest better now
Download I think we need to tune the load on hopeful-rivest better now
10:02:45
@hexa:lossy.networkhexa7 browsers and 3 kernels is a bit much10:03:03
@k900:0upti.meK900 I wonder if less but bigger jobs will be better now 10:03:20
@hexa:lossy.networkhexaRedacted or Malformed Event10:03:24
@hexa:lossy.networkhexahttps://github.com/NixOS/infra/pull/109310:09:46
@hexa:lossy.networkhexa6x20C (down from 10x20C) for aarch64-linux10:10:04
@hexa:lossy.networkhexa6x24C (up from 4x24C) for x86_64-linux10:10:17
@hexa:lossy.networkhexathat's ~1.5 times the core count of the machine10:10:47
@k900:0upti.meK900 Now that we have pressure info, we should really start taking that into account when scheduling 10:10:49
@hexa:lossy.networkhexawe kinda do10:11:01
@k900:0upti.meK900 Would make life a lot easier I think 10:11:03

Show newer messages


Back to Room ListRoom Version: 6