| 19 Sep 2025 |
 lord_fomo | https://nixops.dev/manual/development/ | 19:22:50 |
| lord_fomo | oof | 19:23:14 |
 K900 | I personally don't think it's that interesting, there's just too much critical mass behind Terraform/OpenTofu and nixops4 doesn't significantly improve on that model | 19:24:27 |
| lord_fomo | well having come from that space years ago, i can claim that terraform is not a great solution if you actually scale to large clusters | 19:25:06 |
| lord_fomo | saltstack / ansible would be much better solutions | 19:25:16 |
| lord_fomo | but i guess i was kinda of hoping for a next-gen soln coming from nixland | 19:25:29 |
| K900 | Those don't solve the same problem | 19:25:42 |
| lord_fomo | pretty sure they do | 19:25:53 |
| lord_fomo | but maybe things changed | 19:26:00 |
| lord_fomo | i also know people workin in devops that would verify those claims pretty sure | 19:26:14 |
| K900 | Terraform is not in the same class of thing as Ansible, and it never was | 19:26:18 |
| lord_fomo | sure, then salt | 19:26:28 |
| lord_fomo | i wouldn't use ansible anyway over salt | 19:26:35 |
| lord_fomo | or openstack | 19:26:39 |
| K900 | Salt and Ansible are roughly in the same class of thing | 19:26:40 |
| K900 | Which Terraform is not | 19:26:44 |
| K900 | Openstack is in yet another class of thing that is neither of those two | 19:26:56 |
| lord_fomo | well i know for a fact you can full-lifecycle deploy on salt bc i patched it for that years ago | 19:27:10 |
| K900 | "Full lifecycle" as in "deploy an image and then customize it later", sure | 19:27:59 |
| lord_fomo | on the vmware backend actually of all things, before they were acquired 😂 | 19:27:59 |
| K900 | You can also do that with Terraform + any NixOS deployment tool | 19:28:14 |
| lord_fomo | i mean go from no host to full host | 19:28:23 |
| lord_fomo | presuming some distributed runtime bootstrapping | 19:28:37 |
| K900 | You're generally not going to do better on a public cloud, at least currently | 19:29:32 |
| K900 | And trying to pretend that the two steps are actually the same step doesn't help | 19:29:36 |
| lord_fomo | anyway, i'm slightly surprised you wouldn't be more bullish on a nixland approach actually | 19:29:37 |
| lord_fomo | to me that's like the whole end goal of everything i'd think nix is trying to do? | 19:29:54 |
| lord_fomo | In reply to @k900:0upti.me
You're generally not going to do better on a public cloud, at least currently fair | 19:30:04 |
| K900 | I think a "nixland approach" at this point requires cooperation from the control plane | 19:30:13 |
| K900 | And such a control plane can absolutely exist | 19:30:20 |
