# I migrated my complete homelab from Proxmox to Nix — retained record

> Retained reddit record from reddit:homelab.

Canonical: https://highsignal.app/data/records/7c316e005b3b2b0c

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:homelab)  published 2026-10-24 14:42 ingested 2026-10-25 02:38

[original source ↗](https://www.reddit.com/r/homelab/comments/1xeaf64/i_migrated_my_complete_homelab_from_proxmox_to_nix/)

## retained text



```
I recently migrated from Proxmox to Nix for everything. I use qemu for vms and nspawn for containers. The security model is similar to LXC with guest uid offset and restrictions using systemd but there is no apparmor integration yet. I also tested firecracker earlier but decided to settle on nspawn for now.

The reasons to do this were,

1. Uniform configuration across all services. I used to initialise my proxmox containers with ansible and then do manual configuration but over 7 years a lot of that drifts and some services end up being better configured than others.

2. I was using 3 node proxmox cluster but I was mostly not using any HA features. Proxmox in cluster mode is very harsh on consumer grade SSDs and I think this is fixable to some extent but I didnt want to deal with it. I don't care about zero downtime migrations or distributing traffic k8s style across nodes in a proxmox cluster.

3. With a good config, its easy to have everything better configured. Previously, it would create large backups of the whole volume and proxmox does not create deltas. This wastes a lot of local disk space and it was more expensive for online backups. I could’ve addressed it to some extent by creating more volumes and then only backing up app data volumes rather than the whole service volume but I felt it was annoying to setup. The current version lets me do that very easily. The backups size dropped from around 450G(full volume backups and multiple old copies) to just around 15-20G and its super easy to adjust this now. 

My repo is here: https://github.com/ishanjain28/homelab

Dashboards and status pages:

1.	⁠https://dash.ishanjain.me/public-dashboards/c6bb8550a0314dd9b6f5bfe4b27afc78  

2.	⁠https://dash.ishanjain.me/public-dashboards/8df11f9eb01d433297f691ab50ac704c

3. https://status.ishanjain.me

(2 more nodes are unavailable right now 1 was running some vms and the the other one is not switched over yet.)

Status page config is auto generated from service definition. I have a branch to do the same for with caddy proxy config but I need to figure out a good way to do it without just dumping all the domains in public.

The VM section was fully generated by LLMs. I initially wrote a simpler version that was 1:1 copy of my VMs in proxmox and then an llm pointed out the performance problems and I let it write the whole thing.

Happy to answer any questions!
```



document fetched 2026-10-25 00:32 · canonical https://reddit.com/r/homelab/comments/1xeaf64/i_migrated_my_complete_homelab_from_proxmox_to_nix
