| 9 Mar 2026 |
 hexa | I think you track them quite closely | 23:37:52 |
| 10 Mar 2026 |
 Tom | Well, to be fair: even upstream struggles keeping pace with the speed by which @adam:robins.wtf opens PRs 😂 🙃 | 00:13:07 |
 adamcstephens | atom feeds are nice :) | 02:47:24 |
| adamcstephens | i use miniflux to track releases, and of course other things | 02:47:49 |
| 13 Mar 2026 |
| hexa | any reason why startsWhenNeeded is enabled on openssh in lxc common? | 16:28:01 |
| hexa | that implies socket activation, which uses sshd@ units with instance names containing src/dst address | 16:28:23 |
| hexa | which again land in prometheus :D | 16:28:31 |
| adamcstephens | pretty sure that's a hold over from before i even started with nixos. there's probably a small argument that it's saving memory, but maybe not worth it | 17:23:15 |
| 14 Mar 2026 |
|  Milja K joined the room. | 11:10:56 |
| 15 Mar 2026 |
|  @jykrwn_bot:matrix.org joined the room. | 01:23:27 |
|  NixOS Moderation Bot banned @jykrwn_bot:matrix.org (disagreement). | 20:28:00 |
| 20 Mar 2026 |
| hexa | the newest fun time I have is when the ext4 filesystem of a guest gets corrupted during online migration with zfs send, when the host is busy with i/o | 16:41:46 |
| hexa | in our cast mostly gitlab runner | 16:41:52 |
| hexa | e.g. when building nix things | 16:42:08 |
| hexa | three weeks in a row now | 16:42:31 |
| hexa | Redacted or Malformed Event | 16:45:16 |
| adamcstephens | "fun" | 17:08:24 |
| hexa | yup | 17:26:26 |
| hexa | Which makes it pretty clear that I'm going to stay with ceph for other setups, even if writes are more expensive. | 21:04:51 |
| hexa | Now if only the nixpkgs packaging of ceph was better | 21:05:37 |
| Tom | better in which aspect? Haven't looked at it just heard that ceph and nixos aren't necessarily the best combo. | 21:09:58 |
| hexa | they could be | 21:18:27 |
| hexa | arrow-cpp-19.nix
boost-1.85.patch
boost-1.86-PyModule.patch
old-python-packages/
old-python-packages/cryptography-vectors.nix
old-python-packages/cryptography.nix
old-python-packages/python-cryptography-Cherry-pick-fix-for-CVE-2023-49083-on-cryptography-40.patch
old-python-packages/trustme.nix
| 21:18:42 |
| hexa | nix-repl> ceph.version
"19.2.3"
| 21:19:16 |
| hexa | squid will run out of support during the 26.05 release cycle | 21:19:59 |
| hexa | there is no clear maintainer for this package | 21:20:19 |
| Tom | oh, there are even two PRs for it 😅
https://github.com/NixOS/nixpkgs/pull/494887 https://github.com/NixOS/nixpkgs/pull/494583 | 21:21:49 |
| Tom | the second does look like it improves the overall situation | 21:39:21 |
| Tom | but yeah, there curently doesn't seem to be an (active) committer involved | 21:52:25 |
| 24 Mar 2026 |
| adamcstephens | huh, I didn't realize we'd added an nspawn driver for tests. That could really help the Incus tests since they rely on nested virtualization which is problematic outside x86 | 20:03:10 |
