| 26 May 2025 |
 @alex:acrawford.com | You’re modifying /etc/configuration.nix on the running machine and then doing nixos rebuild switch? | 18:24:48 |
 gotha | yes, exactly | 18:44:42 |
| 27 May 2025 |
|  jsw left the room. | 07:10:47 |
|  @irenes:matrix.org left the room. | 08:59:53 |
|  Lorenz joined the room. | 11:39:28 |
|  @federicodschonborn:matrix.org changed their display name from LEGO® Worm™ (Asbestos Flavored!) to LEGO® Worm™ (Now Only 1 krad!). | 14:42:36 |
|  @unlimitedsola:matrix.org left the room. | 15:45:38 |
|  txrx joined the room. | 19:35:54 |
| txrx | If I have a flake that specifies more than one system, and I want to share a string (eg ssh key, ip address, etc) between configs for multiple systems, is the best thing to create a shared module with something like options.foo = { sharedString = lib.mkOption { ... }; };, and then refer to config.foo.sharedString? I keep second guessing this because I'm specifying what is in my case a constant as a configurable option. | 19:43:06 |
 @kongrooparadox:matrix.org | an option is useful if you are going to override the value when you call the module
for a shared variable, here is what I do with my authorized ssh keys for example.
# ssh.nix
{
keys = [ "pubKey1" "pubKey2"];
}
And you import the keys in your config :
users.users.root.openssh.authorizedKeys.keys = (import ./relative-path/to/ssh.nix).keys;
| 20:56:10 |
| @kongrooparadox:matrix.org | * an option is useful if you are going to override the value when you call the module
for a shared variable, here is what I do with my authorized ssh keys for example.
# ssh.nix
{
keys = [ "pubKey1" "pubKey2" ];
}
And you import the keys in your config :
users.users.root.openssh.authorizedKeys.keys = (import ./relative-path/to/ssh.nix).keys;
| 20:57:42 |
| 28 May 2025 |
|  Sylonin joined the room. | 03:02:01 |
| 29 May 2025 |
|  Adnan Zaman changed their profile picture. | 12:29:23 |
| Adnan Zaman changed their profile picture. | 12:33:14 |
|  raul changed their display name from razterer to raul. | 21:19:00 |
| 30 May 2025 |
|  . changed their display name from adeci to .. | 00:39:16 |
| . changed their profile picture. | 00:39:51 |
|  George joined the room. | 12:14:32 |
|  DerivationDingus set a profile picture. | 19:53:25 |
| DerivationDingus changed their profile picture. | 19:54:41 |
 finga | I just upgraded my server to 25.05 and there I have some containers, one for matrix-synapse, one for xandikos (a webdav server) and one for a nginx proxy which proxies matrix-synapse and xandikos. now since the update the nginx container fails to start, the log is full with "A start job is running for Renew ACME certificate for..." when I stop the acme-setup.service on the nginx container before the container fails due to timeout everything seems fine so far... Is this a known issue? What can I do to fix this? | 23:59:09 |
| 31 May 2025 |
| finga | I forgot to mention, when I remove the vhost for matrix but keep the one for xandikos, everything works fine as expected. | 00:00:30 |
| gotha | if anyone is curious what is going on here, I was running a VM with 1GB of RAM and no swap and it seems like it was running out of memory
nixos-rebuild was just dying without error ..
adding a swapfile fixed it | 10:48:55 |
| gotha | * if anyone is curious what is going on here, I was running a VM with 1GB of RAM and no swap and it seems like it was running out of memory
nixos-rebuild was just dying without error ..
adding a swapfile fixed it
I am not sure if better error message is even possible in this scenario ... | 10:49:30 |
 K900 | Yeah | 10:49:32 |
| K900 | It's not really possible | 10:49:36 |
| K900 | But you should see an OOM kill in dmesg | 10:49:40 |
| K900 | A nixpkgs eval needs ~3GB of RAM | 10:49:44 |
|  chipset joined the room. | 13:23:40 |
|  Rajiv Singh joined the room. | 17:10:57 |
