| 24 May 2024 |
 raitobezarius | It's a bit clunky backward-compatible wise | 17:23:53 |
| raitobezarius | I would need to test more scenarios and would recommend making deployment.systemType mandatory on the long term, etc. | 17:24:05 |
| raitobezarius | but here's an example: https://git.dgnum.eu/DGNum/infrastructure/src/branch/colmena-liminix/hive.nix where both system types areu sed | 17:24:26 |
| raitobezarius | * but here's an example: https://git.dgnum.eu/DGNum/infrastructure/src/branch/colmena-liminix/hive.nix where both system types are used | 17:24:28 |
| raitobezarius | NixOS ones and Liminix ones | 17:24:31 |
| raitobezarius | it's a bit difficult to read our machinery but: https://git.dgnum.eu/DGNum/infrastructure/src/branch/colmena-liminix/meta/nodes.nix#L90-L95 mention the Liminix node and get massaged into https://git.dgnum.eu/DGNum/infrastructure/src/branch/colmena-liminix/hive.nix#L33-L36 | 17:25:16 |
| raitobezarius | I am also trying to wonder how to move the networking part inside of the Hive expressions | 17:26:23 |
| raitobezarius | And let user define new networking mechanisms | 17:26:27 |
| raitobezarius | Usecase: deploy over serial console | 17:26:32 |
| raitobezarius | (or deploy over SSH3) | 17:26:48 |
| raitobezarius | I thought of having something like NixDefined implementation of Host | 17:27:33 |
| raitobezarius | that gets in input some scripts that can perform all the trait operations of Host | 17:27:44 |
| raitobezarius | which gets passed via Nix | 17:27:51 |
 Zhaofeng Li | Hmm, that could also be precursor to supporting alternative deployment models (pull-based, some kind of central dispatch server, etc.) | 17:29:21 |
| raitobezarius | possibly | 17:30:17 |
| raitobezarius | I admit that the design space is so immense that it's hard to be too generic | 17:30:27 |
| Zhaofeng Li | Indeed, we have to be opinionated in many cases | 17:50:59 |
| 29 May 2024 |
| raitobezarius changed their display name from raitobezarius to raitobezarius (DECT: 7248). | 17:09:53 |
| 30 May 2024 |
|  Unit 1721344 [polygon] (it/its, es/ihr) changed their display name from polygon system (she/it/they) to polygon system (she/it/they) [DECT 7659]. | 09:31:28 |
|  Tammi (ey/em) left the room. | 15:42:19 |
| 4 Jun 2024 |
| raitobezarius changed their display name from raitobezarius (DECT: 7248) to raitobezarius. | 11:15:55 |
| 5 Jun 2024 |
| Unit 1721344 [polygon] (it/its, es/ihr) changed their display name from polygon system (she/it/they) [DECT 7659] to polygon system (it/she/they). | 11:57:33 |
|  f0x joined the room. | 13:18:43 |
| f0x | while reading through the Colmena docs on secrets, it notes
Take note that if you use the default path (/run/keys), the secret files are only stored in-memory and will not survive reboots
am I missing something or would you basically always want to use a different path then?
| 13:33:03 |
 Tom | depends on your usage scenario, uploading secrets only to RAM has security benefits | 13:39:45 |
| f0x | right, but you'd need to (manually) deploy them again every time you boot | 13:41:41 |
| Tom | sure, but for servers that might not be a problem or even desired | 13:42:24 |
| Tom | * sure, but for servers that might not be a problem or is even desired | 13:47:14 |
