!RROtHmAaQIkiJzJZZE:nixos.org

NixOS Infrastructure

389 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.120 Servers

Load older messages


SenderMessageTime
4 Oct 2021
@K900:matrix.orgK900Which loops forever13:19:56
@K900:matrix.orgK900This is using Cloudflare's name servers so shouldn't be ISP fuckery13:20:43
@janne.hess:helsinki-systems.deJanne Heß left the room.13:27:10
@zimbatm:numtide.comJonas Chevaliertracking here: https://github.com/NixOS/nixos-org-configurations/issues/17713:35:18
@zimbatm:numtide.comJonas Chevalierboth the old A record and the new NETLIFY records were living side-by-side. This should be fixed now once DNS propagates (~1h)13:40:51
@K900:matrix.orgK900Thanks, seems good now :) 13:54:13
@baughn:brage.infoBaughn joined the room.19:52:20
@baughn:brage.infoBaughnThe iso files for the gnome desktop, as well as both minimal installation images on nixos.org don't match their sha256sums.19:52:51
@vcunat:matrix.orgVladimír ČunátThe sums are for a different commit than the actual ISOs.19:56:56
@vcunat:matrix.orgVladimír Čunát For minimal I see:
21.05.3509.7daf35532d2
21.05.3468.92609f3d9bc
19:57:39
@baughn:brage.infoBaughnMmyep. There's the bug.19:58:34
@vcunat:matrix.orgVladimír ČunátI wonder if it's possible that these redirect pairs aren't guaranteed to be "cached atomically" by Fastly or something.19:59:41
@linus.heckemann:matrix.mayflower.deLinux Hackerman joined the room.19:59:42
@djacu:matrix.orgdjacu joined the room.19:59:51
@vcunat:matrix.orgVladimír Čunát * I wonder if it's possible that these redirect pairs aren't guaranteed to be "cached atomically" by Fastly or something. (well, I don't know this stuff)20:00:08
@baughn:brage.infoBaughnI'd call that virtually guaranteed20:00:12
@baughn:brage.infoBaughnAs in, guaranteed not to work20:00:22
@baughn:brage.infoBaughnWe'll find out in a couple minutes.20:00:27
@baughn:brage.infoBaughnAn ideal fix would be to have a JSON file on the website which points to the iso, and also contains the sha256 of that iso. That's cover the browser case, at least; for the console case, you could have a single directory that's a redirect to a latest-iso dir containing both files20:01:41
@baughn:brage.infoBaughn * An ideal fix would be to have a JSON file on the website which points to the iso, and also contains the sha256 of that iso. That'd cover the browser case, at least; for the console case, you could have a single directory that's a redirect to a latest-iso dir containing both files20:01:48
@linus.heckemann:matrix.mayflower.deLinux HackermanThat's one option. Surely the CDN should have a way to invalidate caches though?20:02:11
@baughn:brage.infoBaughnProbably, but don't go that path. Cache invalidation is the hardest problem in computer science.20:02:56
@baughn:brage.infoBaughn Much, much better to use a construction that makes it unnecessary. 20:03:05
@linus.heckemann:matrix.mayflower.deLinux HackermanTrue20:04:39
@baughn:brage.infoBaughnI'd be happy to supply the PRs to fix this, honestly... dunno where I should look, though!20:05:32
@sushi_dude:matrix.orgSushi Dude joined the room.20:07:18
@baughn:brage.infoBaughnFound it~20:18:25
5 Oct 2021
@k900:0upti.meK900 joined the room.07:06:28
@K900:matrix.orgK900 left the room.07:15:53
6 Oct 2021
@rosariopulella:matrix.orgRosario Pulella changed their display name from rosariopulella to Rosuavio.10:38:42

Show newer messages


Back to Room ListRoom Version: 6