!RROtHmAaQIkiJzJZZE:nixos.org

NixOS Infrastructure

476 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
18 Jun 2026
@arianvp:matrix.orgArian
In reply to @joerg:thalheim.io
You can't really stream zstd with s3 unfortunally because s3 wants to know the size up-front. In niks3 I solved this by having smallish multi-part upload. I just yesterday introduced something similar in queue-runner because before it was just buffering everything in memory.
S3 doesn't require this
16:14:18
@arianvp:matrix.orgArianOr maybe I'm misunderstanding what we mean with size upfront16:14:49
@xokdvium:matrix.orgSergei Zimmerman (xokdvium) I mean, if we want to be able to retry/send the hashsums as x-amz- headers and stuff yeah, it’s kinda necessary to keep the file around.
In nix we that by compressing to an O_TMPFILE fd and seeking in it when we need to retry
16:17:10
@joerg:thalheim.ioMic92AWS maybe, but there are also other s3 compatible endpoints that require 'Content-Length' headers to be set.16:35:12
@arianvp:matrix.orgArianAh wait maybe you're right. NVM 17:33:43
@arianvp:matrix.orgArianExcept for multipart uploads but we don't need those for small objects 17:33:59
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)
In reply to @arianvp:matrix.org
Ah wait maybe you're right. NVM
I don’t think we set Content-Length in nix’s s3 client?
17:35:01
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)Just for POST, not for PUT17:35:37
@xokdvium:matrix.orgSergei Zimmerman (xokdvium)Ah no, sorry. We did in fact set it for both17:36:22
@hexa:lossy.networkhexahttps://grafana.nixos.org/d/hejbnzw/iocaine17:44:30
@hexa:lossy.networkhexahydra feels much more responsive when it is not hammered by bots17:48:09
@hexa:lossy.networkhexathat manage to bypass anubis17:48:14
@magic_rb:matrix.redalder.orgmagic_rbDid you swap anubis with iocaine?17:48:43
@hexa:lossy.networkhexalooks like unstable-small just advanced18:04:48
@whispers:catgirl.cloudwhispers [& it/fae] changed their display name from whispers [& it/fae] to meow meow.18:46:02
@whispers:catgirl.cloudwhispers [& it/fae] changed their display name from meow meow to whispers [& it/fae].19:11:49
@stv0ge:matrix.orgstv0g joined the room.19:14:02
@graysontinker:matrix.orgGrayson Tinker joined the room.19:31:49
@joerg:thalheim.ioMic92Looks I got it fixed now with my parallel upload code: https://hydra.nixos.org/machines all machines are building and rarely upload19:59:15
@joerg:thalheim.ioMic92* Looks I got it fixed now with the new parallel upload code: https://hydra.nixos.org/machines all machines are building and rarely upload19:59:22
@joerg:thalheim.ioMic92* Looks I got it fixed now with the new parallel upload code: https://hydra.nixos.org/machines all machines are building and rarely uploading19:59:35
@tobias:r0.ggTobi joined the room.20:25:50
19 Jun 2026
@vcunat:matrix.orgVladimír ČunátSame derivation is being built on two machines at once? At a glance, this happens even in a situation where the previous queue-runner wouldn't do it: https://hydra.nixos.org/build/331659169#tabs-buildsteps03:28:10
@joerg:thalheim.ioMic92Did the first build actually finish?05:31:00
@joerg:thalheim.ioMic92Maybe just a missing DB update 05:31:27
@k900:0upti.meK900 https://hydra.nixos.org/build/331675357 another funni 05:38:13
@joerg:thalheim.ioMic92I just checked just looks like stale db entries05:48:11
@joerg:thalheim.ioMic92There are not actually two builds running05:48:27
@joerg:thalheim.ioMic92Looks the db isn't properly updated05:49:16
@vcunat:matrix.orgVladimír ČunátThe first one did not run until the end.05:49:20

Show newer messages


Back to Room ListRoom Version: 6