!RROtHmAaQIkiJzJZZE:nixos.org

NixOS Infrastructure

373 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.114 Servers

Load older messages


SenderMessageTime
26 Jul 2025
@k900:0upti.meK900I assume no19:02:32
@vcunat:matrix.orgvcunatNo, just the "blob data" lines.19:03:00
@vcunat:matrix.orgvcunatBut I assume the indexing part is relatively easy to reproduce anywhere.19:03:46
@k900:0upti.meK900Unfortunately it's very network bound19:04:54
@k900:0upti.meK900In most cases19:04:56
@vcunat:matrix.orgvcunatGrafana shows about half a core being busy during the indexing.19:08:24
@vcunat:matrix.orgvcunatFrom that I'd roughly estimate that waiting for network roughly doubled the real time?19:10:15
@vcunat:matrix.orgvcunat* From that I'd estimate that waiting for network roughly doubled the real time?19:10:37
@vcunat:matrix.orgvcunatAnyway, I don't see practical problems from running ~20 minutes.19:11:06
@vcunat:matrix.orgvcunat* Anyway, I don't see practical problems from this taking ~20 minutes.19:11:15
@k900:0upti.meK900I don't either19:12:08
@k900:0upti.meK900But it is abnormal19:12:11
@vcunat:matrix.orgvcunatEven if we're talking about some critical security issues, this step will still be just a small fraction of the whole update process.19:12:19
@k900:0upti.meK900It's not that I think it's an issue overall19:13:03
@k900:0upti.meK900But when a thing suddenly becomes 5x slower19:13:10
@k900:0upti.meK900That's usually a sign19:13:14
@k900:0upti.meK900 Of something 19:13:19
@vcunat:matrix.orgvcunatOther -small updates today seemed faster, at a quick glance.19:13:30
@vcunat:matrix.orgvcunat 15 minutes for the previous unstable-small less than a day ago. 19:16:47
@hexa:lossy.networkhexa (signing key rotation when)imo things started taking longer since the debuginfo changes22:59:34
@hexa:lossy.networkhexa (signing key rotation when)but that is a subjective observation22:59:49
27 Jul 2025
@idabzo:matrix.org@idabzo:matrix.org left the room.10:28:31
@wamserma:nixos.devwamsermathere seems to be some issue with the aarch64-linux builders: https://hydra.nixos.org/build/30291170021:12:24
@hexa:lossy.networkhexa (signing key rotation when)
  UID     PID    PPID STIME     TIME COMMAND
30001  423172  423139 08:15 00:00:00 bash -e /nix/store/vj1c3wf9c11a0qs6p3ymfvrnsdgsdcbq-source-stdenv.sh /nix/store/shkw4qm9qcw5sc5n1k5jznc83ny02r39-default-builder.sh
30001  423217  423172 08:15 00:00:00 make SHELL=/nix/store/7z7zi55hb1kck2sm4jx0sbcbwcmajyrh-bash-5.3p0/bin/bash PREFIX=/nix/store/2gfard3ywqs2g13525qwfzbxr4zl6rji-owl-lisp-0.2.2 CC=cc
30001  429283  423217 08:16 02:04:10 sh tests/run all bin/olp
21:16:02
@hexa:lossy.networkhexa (signing key rotation when)stuck in the tests21:16:19
@hexa:lossy.networkhexa (signing key rotation when)killed21:16:40
@hexa:lossy.networkhexa (signing key rotation when)image.png
Download image.png
21:16:55
@hexa:lossy.networkhexa (signing key rotation when)could be a regression in the last staging cycle21:17:30
@hexa:lossy.networkhexa (signing key rotation when)image.png
Download image.png
21:17:41
@hexa:lossy.networkhexa (signing key rotation when)Redacted or Malformed Event21:18:02

Show newer messages


Back to Room ListRoom Version: 6