| 19 Sep 2025 |
 dgrig | That would be very helpful if I'm understanding the issue correctly, thank you! | 12:08:30 |
| dgrig | A bit off-topic for the room, but related to nix-eval-jobs. If you know what's up with clang-tidy here I would love a hint. | 12:12:17 |
 ghpzin | I remember somebody here mentioning problems with https://releases.nixos.org a few days ago that got fixed.
I noticed something strange is happening with nixos/unstable,nixos/unstable-small and nixpkgs listings:
https://releases.nixos.org/?prefix=nixos/unstable/
https://releases.nixos.org/?prefix=nixos/unstable-small/
https://releases.nixos.org/?prefix=nixpkgs/
Page keeps sending GET requests (that look like https://nix-releases.s3.amazonaws.com/?delimiter=/&prefix=nixos/unstable/&marker=nixos%2Funstable%2Fnixos-21.03pre248308.5d0e2dedd55%2F) over and over again and not showing anything on page.
Other pages seem to list things fine and not spam requests. | 16:35:24 |
 Mic92 | @ghpzin:envs.net: https://github.com/NixOS/infra/pull/839 related to this? | 18:53:37 |
| ghpzin | Yes, that was for previous issue when there was js error and nothing loaded.
Not sure if related to security fix that was the reason for that. | 18:58:33 |
| ghpzin | * Yes, that was for previous issue when there was js error and nothing loaded.
Not sure if current issue is related to security fix that was the reason for that. | 19:00:58 |
| ghpzin | Yes, that was for previous issue when there was js error and nothing loaded.
Not sure if current issue is related to security fix that was the reason for that.
It is weird, because on other prefixes there is only 1 request and it is fine, on those there is first one and then the always spamming one.
Both return 200 with what seems to be correct things: <ListBucketResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Name>... | 19:04:20 |
| ghpzin | Yes, that was for previous issue when there was js error and nothing loaded.
Not sure if current issue is related to security fix that was the reason for that.
It is weird, because on other prefixes there is only 1 request and it is fine, on those there is first one and then the always spamming one.
Both return 200 with what seems to be correct things: <ListBucketResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Name>...
I guess pagination does not work, first request on spamming ones returns:
<MaxKeys>1000</MaxKeys><Delimiter>/</Delimiter><IsTruncated>true</IsTruncated>
Working ones return <IsTruncated>false</IsTruncated> | 19:10:23 |
| ghpzin | Yes, that was for previous issue when there was js error and nothing loaded.
Not sure if current issue is related to security fix that was the reason for that.
It is weird, because on other prefixes there is only 1 request and it is fine, on those there is first one and then the always spamming one.
Both return 200 with what seems to be correct things: <ListBucketResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Name>...
I guess pagination does not work, first request on spamming ones returns:
<MaxKeys>1000</MaxKeys><Delimiter>/</Delimiter><IsTruncated>true</IsTruncated>
Working ones return <MaxKeys>1000</MaxKeys><Delimiter>/</Delimiter><IsTruncated>false</IsTruncated> | 19:10:51 |
| ghpzin | Yes, that was for previous issue when there was js error and nothing loaded.
Not sure if current issue is related to security fix that was the reason for that.
It is weird, because on other prefixes there is only 1 request and it is fine, on those there is first one and then the always spamming one.
Both return 200 with what seems to be correct things: <ListBucketResult xmlns="http://s3.amazonaws.com/doc/2006-03-01/"><Name>...
I guess pagination does not work, first request on spamming ones returns:
<IsTruncated>true</IsTruncated>
Working ones return <IsTruncated>false</IsTruncated> | 19:11:19 |
 vcunat | I think those pages did work at some point, but I haven't tried very recently to be sure that they did break right now around those security fixes. | 19:14:24 |
| vcunat | * I believe those pages did work at some point, but I haven't tried very recently to be sure that they did break right now around those security fixes. | 19:14:32 |
 K900 | Can someone update anubis on Hydra | 19:14:40 |
| K900 | I still can't load it on my phonen | 19:14:46 |
| K900 | * | 19:14:50 |
| K900 | Because of the 9 cores thing | 19:14:55 |
| vcunat | We deploy whatever is in stable-small. | 19:16:28 |
| vcunat | I think even in this case (anubis), at a quick glance at the repo. | 19:16:51 |
| vcunat | Last update was just a couple days ago. | 19:17:04 |
| K900 | Hm I wonder if they haven't tagged a release yet | 19:18:27 |
| Mic92 | What is the 9 cores thing? | 20:19:28 |
| dgrig | https://anubis.techaro.lol/blog/2025/cpu-core-odd/ | 20:25:13 |
| ghpzin | Commit with fix seems to be only in v1.22.0:
https://github.com/TecharoHQ/anubis/commit/c661bc37d133976a005fb0b1060e3e0a34f69003 | 20:28:20 |
| 21 Sep 2025 |
 hexa (signing key rotation when) | getting closer to replacing haumea | 21:25:36 |
| hexa (signing key rotation when) | but the requirements changed slightly | 21:25:47 |
| hexa (signing key rotation when) | given the recent increase of low disk alerts we should go for 2x2TB instead of 2x1TB | 21:26:03 |
| hexa (signing key rotation when) | upgrading an AX52 will land us at 124,63 EUR/mo | 21:27:05 |
| hexa (signing key rotation when) | an AX102 starts at 125,84 EUR/mo and comes with twice the memory and a beefier CPU | 21:27:43 |
| hexa (signing key rotation when) | the intel alternative would be an EX101 but I'm not fond of the prevalence of E-cores on the i9-13900 | 21:28:18 |
| hexa (signing key rotation when) | https://www.hetzner.com/dedicated-rootserver/matrix-ax/
https://www.hetzner.com/dedicated-rootserver/matrix-ex/ | 21:28:34 |
