Fix LXC deployment path and clean up remaining eval warnings

LXC hosts (device busy fix):

modules/platforms/lxc.nix now imports nixpkgs' own
virtualisation/proxmox-lxc.nix, giving every lxc-* host a real
config.system.build.tarball output — a directly `pct restore`-able
Proxmox container image. This is the actual bug fix behind the
"cannot remove real root directory: device busy or in use" error:
lxc-* targets were only reachable through nixos-install, which
bind-mounts / onto /mnt for containers (no raw disk to partition)
and then correctly refuses to modify the filesystem it's currently
running on. auto-install.sh's menu now excludes lxc-* targets
entirely (they deploy via nix build + pct restore instead, see
docs/auto-installer.md) — and, on the same reasoning, also excludes
`installer`/`proxmox-lxc`, which are the installer image's own flake
targets, not deployable hosts.

manageHostName = true keeps host.nix's declared hostnames (upstream's
default would let Proxmox's ambient container config win instead);
privileged = false matches how these containers are actually created.

Eval warnings, now zero across all 19 nixosConfigurations + 4 packages:

- Multiple password options (root/nixos in the installer): nixpkgs'
  own installer profile sets initialHashedPassword = "" for
  passwordless login, conflicting with our explicit hashedPassword.
  Force-nulled the upstream option rather than adopting passwordless
  login, since this image now also boots over LAN PXE with
  PasswordAuthentication enabled.
- boot.zfs.forceImportRoot default value: set explicitly to false
  (matching the two places that already did) in
  modules/common/configuration.nix and modules/installer/common.nix,
  covering every host and the installer alike.
- Deprecated pkgs.system in modules/build-types/gui.nix: switched to
  pkgs.stdenv.hostPlatform.system.

All confirmed non-behavioral where it matters: unrelated hosts'
drvPaths are byte-identical to their pre-existing baselines throughout.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot

docs/auto-installer.md

@@ -12,6 +12,38 @@ discovers available hosts from this same flake, lets the operator choose a
 target, applies that host's Disko storage configuration, installs NixOS, and
 reboots.
 
+**This applies to every `nixosConfigurations` target except `lxc-*` hosts —
+see "LXC hosts" immediately below for why those are different.**
+
+## LXC hosts
+
+`lxc-*` targets (`lxc-minimal`, `lxc-nix-cache`, `lxc-server`, `lxc-docker`,
+`lxc-gui`, `lxc-pxe-boot`) are **not** installed via `auto-install.sh` — the
+interactive menu deliberately excludes them. Don't try to select one there;
+`nixos-install` would bind-mount `/` onto `/mnt` (LXC containers have no raw
+disk to partition) and then refuse to touch the filesystem it's currently
+running on — it's designed to protect exactly this case, so it just fails.
+
+`modules/platforms/lxc.nix` imports nixpkgs' own
+`virtualisation/proxmox-lxc.nix` module, which gives every `lxc-*` host a
+`config.system.build.tarball` output — a complete, directly Proxmox-importable
+container image, no install step at all:
+
+```sh
+nix build .#nixosConfigurations.lxc-minimal.config.system.build.tarball
+```
+
+Then, on the Proxmox host, `pct restore` (or the GUI's "Create CT" → upload
+template flow) that tarball directly as a new container. First boot runs
+`boot.postBootCommands` (registers the Nix store DB and system profile) —
+there's no separate activation step to run yourself.
+
+Host keys still need pre-seeding the same way as any other host (see "Host
+keys" below) — the sops-nix activation-vs-first-boot race is identical
+regardless of how the image reaches the machine. `NIXOS_HOST_KEYS_DIR=...
+nix build ... --impure` bakes the matching key into the tarball the same way
+it does for the ISO/PXE installer images.
+
 ## Layout
 
 - `modules/installer/common.nix` — shared by every installer target: SSH

modules/build-types/gui.nix

@@ -2,7 +2,7 @@
 
 {
   environment.systemPackages = with pkgs; [
-    inputs.nixos-conf-editor.packages.${pkgs.system}.nixos-conf-editor
+    inputs.nixos-conf-editor.packages.${pkgs.stdenv.hostPlatform.system}.nixos-conf-editor
     nodejs
     appimage-run
     seahorse

modules/common/configuration.nix

@@ -13,6 +13,12 @@
 
   networking.networkmanager.enable = true; # Easiest to use and most distros use this by default.
 
+  # Recommended over the true default (bypasses ZFS's own import safeguards)
+  # per the option's own docs; matches hosts/docker/host.nix and
+  # modules/services/zfs/enable-service.nix, which already set this
+  # explicitly. Harmless no-op on hosts that don't use ZFS at all.
+  boot.zfs.forceImportRoot = false;
+
   # Set your time zone.
   time.timeZone = vars.timeZone;
 

modules/installer/common.nix

@@ -7,6 +7,12 @@
 
   networking.useDHCP = lib.mkDefault true;
 
+  # Recommended over the true default (bypasses ZFS's own import safeguards)
+  # per the option's own docs. This installer environment has no ZFS pools
+  # of its own to import, so this is a no-op here — just silences the
+  # eval-time warning, matching modules/common/configuration.nix.
+  boot.zfs.forceImportRoot = false;
+
   time.timeZone = vars.timeZone;
 
   # Without this, the installer only ever sees cache.nixos.org, which
@@ -50,11 +56,26 @@
         export FLAKE_BASE_URL="git+https://${vars.lanDomain}/beatzaplenty/nixos.git"
 
         echo "Fetching available NixOS hosts from flake..."
+        # Two categories deliberately excluded from the menu:
+        #   lxc-*                 — these build a config.system.build.tarball
+        #                           meant for `pct restore` on Proxmox
+        #                           directly, not an install. Running
+        #                           nixos-install against one here would
+        #                           bind-mount / onto /mnt and then refuse to
+        #                           touch the filesystem it's currently
+        #                           running on — see docs/auto-installer.md.
+        #   installer/proxmox-lxc — these *are* the installer image's own
+        #                           flake targets, not deployable hosts;
+        #                           "installing" one means nixos-install-ing
+        #                           a copy of the installer into itself.
         mapfile -t options < <(
           nix eval --json --no-use-registries --no-accept-flake-config --extra-experimental-features "flakes nix-command" \
             "''${FLAKE_BASE_URL}#nixosConfigurations" \
             --apply builtins.attrNames \
-          | jq -r '.[]'
+          | jq -r '.[]
+              | select(startswith("lxc-") | not)
+              | select(. != "installer")
+              | select(. != "proxmox-lxc")'
         )
 
         if [[ ''${#options[@]} -eq 0 ]]; then
@@ -62,6 +83,10 @@
           exit 1
         fi
 
+        echo "Note: lxc-* targets aren't installed this way — build them with"
+        echo "  nix build .#nixosConfigurations.<name>.config.system.build.tarball"
+        echo "and 'pct restore' the result on Proxmox directly. See docs/auto-installer.md."
+
         echo "Choose the flake profile to install:"
         select choice in "''${options[@]}"; do
           if [[ -n "$choice" ]]; then
@@ -190,9 +215,18 @@
     PasswordAuthentication = true;
   };
 
+  # nixpkgs' own installer profile (profiles/installation-device.nix, pulled
+  # in via installation-cd-minimal.nix) sets initialHashedPassword = "" for
+  # both users — its own passwordless-login convention for install media.
+  # That's a second, non-null password option alongside our hashedPassword
+  # below, which NixOS warns about as ambiguous precedence. Force it null
+  # rather than adopting passwordless login: this image now also boots over
+  # LAN PXE with PasswordAuthentication enabled, so passwordless root SSH
+  # would be reachable by anyone on the LAN, not just local console.
   users.users.root = {
     hashedPassword =
       "$6$Kwv9KAyvcurAViQF$H4.u3feqGE7lVoNgkFXhE3n2Pmo//9JYDTCz8ifrVHBxPjwa1xMby7tEZ8Bpt5MXs9Rkx6/YbZWxs5CpH0s/70";
+    initialHashedPassword = lib.mkForce null;
   };
 
   users.users.${vars.primaryUser} = {
@@ -206,6 +240,7 @@
 
     hashedPassword =
       "$6$Kwv9KAyvcurAViQF$H4.u3feqGE7lVoNgkFXhE3n2Pmo//9JYDTCz8ifrVHBxPjwa1xMby7tEZ8Bpt5MXs9Rkx6/YbZWxs5CpH0s/70";
+    initialHashedPassword = lib.mkForce null;
 
     openssh.authorizedKeys.keys = [
       vars.adminSshKey

modules/platforms/lxc.nix

@@ -1,36 +1,38 @@
-{ lib, ... }:
+{ lib, modulesPath, ... }:
 
 {
-  boot = {
+  # LXC containers share the host kernel — Proxmox starts them by exec'ing
-    isContainer = true;
+  # /sbin/init directly, no bootloader/initrd involved — and Proxmox has its
+  # own container hostname/network provisioning outside Nix. nixpkgs' own
+  # virtualisation/proxmox-lxc.nix module already handles all of this
+  # correctly (boot.isContainer, loader.initScript, systemd-networkd) and,
+  # critically, provides config.system.build.tarball — a directly
+  # `pct restore`-able container image, no nixos-install/bind-mount needed
+  # (nixos-install refuses to touch the filesystem it's currently running
+  # on, which is exactly what bind-mounting / onto /mnt for an installer
+  # LXC container does).
+  imports = [
+    (modulesPath + "/virtualisation/proxmox-lxc.nix")
+  ];
 
-    loader = {
+  proxmoxLXC = {
+    # host.nix declares each host's real hostname (networking.hostName);
+    # keep that instead of letting Proxmox's ambient container config win.
+    manageHostName = true;
+    # Unprivileged matches how these containers are actually created.
+    privileged = false;
+  };
+
+  boot.loader = {
     grub.enable = false;
     systemd-boot.enable = false;
-
-      # LXC containers share the host kernel — Proxmox starts them by exec'ing
-      # /sbin/init directly, no bootloader/initrd involved. Without this, that
-      # file doesn't correctly launch the current generation, so even a
-      # correctly-installed system can fail to come up after reboot.
-      initScript.enable = true;
-    };
   };
 
-  networking = {
+  # NetworkManager depends on a running udevd to enumerate/classify devices,
-    # boot.isContainer disables services.udev (see nixpkgs'
+  # which boot.isContainer disables (see nixpkgs' container-config.nix) —
-    # virtualisation/container-config.nix), and NetworkManager depends on a
+  # that's what broke DHCP-hostname registration in Pi-hole. The imported
-    # running udevd to enumerate/classify devices. That leaves NM unable to
+  # proxmox-lxc.nix module already switches networking to systemd-networkd
-    # reliably manage the container's veth interface, which is what broke
+  # for the same reason; it just doesn't disable NetworkManager itself,
-    # DHCP-hostname registration in Pi-hole. systemd-networkd talks to the
+  # which modules/common/configuration.nix enables for every host.
-    # kernel over rtnetlink directly and doesn't have that dependency.
+  networking.networkmanager.enable = lib.mkForce false;
-    networkmanager.enable = lib.mkForce false;
-    useNetworkd = true;
-
-    # container-config.nix also defaults this to true, which assumes a
-    # systemd-nspawn-style host bind-mount of /etc/resolv.conf. Real Proxmox
-    # LXC doesn't do that (nixpkgs' own virtualisation/proxmox-lxc.nix module
-    # forces this false for the same reason), so leaving it true silently
-    # breaks DNS instead of falling back to networkd/DHCP-provided servers.
-    useHostResolvConf = lib.mkForce false;
-  };
 }