!RROtHmAaQIkiJzJZZE:nixos.org

NixOS Infrastructure

478 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
5 Jul 2026
@flokli:matrix.orgfloklithey're not deterministic, got bitten by this already.15:04:27
@flokli:matrix.orgfloklior well, they're only deterministic if you never change the tooling to build them.15:05:00
@flokli:matrix.orgfloklihttps://git.snix.dev/snix/snix/commit/180bfc4ce41ad25016aae2e3eb4e7af8c3d185ac15:05:51
@jkarlson:kapsi.fiEmil Thorsøeyeah that was the original premise, so definitely unworkable then, obviously kind of just asking15:05:53
6 Jul 2026
@vcunat:matrix.orgVladimír Čunát

The queue-runner config like this

maxSilentTime = 4 * 3600;

doesn't work. Thunderbird still times out in much less than 4h total (from that silent < 2h)

11:43:28
@vcunat:matrix.orgVladimír Čunát *

The queue-runner config like this

maxSilentTime = 4 * 3600;

doesn't work. Thunderbird still times out on aarch64-linux in much less than 4h total (from that silent < 2h)

11:44:10
@vcunat:matrix.orgVladimír ČunátAlways. Retrying won't help, so it's quite unpleasant.11:44:35
@hexa:lossy.networkhexaRestarted the queue-builder on goofy-hopcroft annd11:51:14
@hexa:lossy.networkhexaRedacted or Malformed Event11:51:15
@hexa:lossy.networkhexa
Download
11:51:34
@hexa:lossy.networkhexasudenly it pucks up new builds again 🤔11:51:44
@joerg:thalheim.ioMic92 hexa (signing key rotation when): I'll have a look this evening again. I have roughly the idea what the issue is I think. I just was having a hydra-free weekend. 11:54:50
@hexa:lossy.networkhexatotally fine, the velocity is decent11:55:08
@hexa:lossy.networkhexahope you enjoyed your time off11:55:15
@hexa:lossy.networkhexaRedacted or Malformed Event11:55:32
@vcunat:matrix.orgVladimír ČunátI think the throughput is noticeably higher than with the old queue-runner already.11:55:40
@vcunat:matrix.orgVladimír ČunátAnd we also dropped x86_64-darwin on 26.11, so that helps.11:56:21
@vcunat:matrix.orgVladimír Čunát * And we also dropped x86_64-darwin on 26.11, so that helps as well. 11:56:25
@hexa:lossy.networkhexayeah, it looks like aarch64-linux is the slowest platform now on unstable11:56:40
@hexa:lossy.networkhexagiven the package count and compute we have11:57:00
@vcunat:matrix.orgVladimír ČunátThe big-parallel sub-queue in particular.11:58:22
@vcunat:matrix.orgVladimír Čunát* The big-parallel "sub-platform" in particular.11:58:30
@vcunat:matrix.orgVladimír Čunát* The big-parallel aarch64-linux "sub-platform" in particular.11:58:46
@joerg:thalheim.ioMic92I had fun with making nix build logs nicer: https://buildbot.dse.in.tum.de/repos/github/TUM-DSE/doctor-cluster-config/builds/78/logs/x86_64-linux.nixos-ruby12:02:07
@hexa:lossy.networkhexacute12:04:34
@vcunat:matrix.orgVladimír Čunát"time off" 😁12:04:34
@hexa:lossy.networkhexaI would still probably want to expand that to have more height12:04:57
@vcunat:matrix.orgVladimír Čunát* "hydra-free" "time off" 😁12:05:01
@joerg:thalheim.ioMic92Yeah. I also noticed or having a button doing a fullscreen increase12:05:53
@hexa:lossy.networkhexayes, the hydra frontend is unresponsive right now20:59:45

Show newer messages


Back to Room ListRoom Version: 6