Distributed builds failed with "Host key verification failed" on any
client that had never manually SSH'd to nix-cache before, since
nothing populated root's known_hosts for it. Wire nix-cache's host
public key into programs.ssh.knownHosts via a new vars.nixCacheHostKey
so every client picks it up automatically on rebuild.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
lxc-* hosts need NIXOS_HOST_KEYS_DIR + --impure to bake in a pre-seeded
SSH host key, otherwise sops-nix's .sops.yaml recipient never matches
and every secret permanently fails to decrypt on first boot. That
invocation is easy to forget, so wrap it as `buildImage <flake-target>`
alongside the existing Switch-nix/Test-nix helpers.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
services.beszel.agent runs under DynamicUser=true with ProtectSystem =
"strict" and no StateDirectory, so /var/lib/beszel-agent -- where the
agent persists the fingerprint that locks its hub pairing to this
machine (github.com/henrygd/beszel/discussions/1542) -- was never
actually writable. Every restart silently failed to persist it and
regenerated a fresh one in memory, permanently desyncing from whatever
the hub had on record after the very first successful pairing. Affects
every host importing modules/beszel/enable-agent.nix (nix-cache, server),
not just full container rebuilds.
Found via nix-cache showing "fingerprint mismatch" after being rebuilt
post-outage; confirmed server was silently exposed to the same bug, just
hadn't restarted since its first pairing. Fixed by declaring
StateDirectory so systemd gives the dynamic user real persistent storage.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
New build type dedicated to Tailscale exit-node capability, wired up for
linode/proxmox/lxc like every other build type (the lxc variant is the one
actually intended for deployment). Kept separate from the "server" host
rather than bundling exit-node capability onto it.
Trimmed modules/tailscale/exit-node.nix down to pure exit-node behavior:
dropped the old --advertise-routes=${vars.lanCidr} bundling (meaningless
for a Linode-hosted VPS with no path to the LAN), and switched
extraUpFlags -> extraSetFlags. Confirmed against nixpkgs' tailscale.nix
that extraUpFlags is only applied by tailscaled-autoconnect, which itself
only runs when services.tailscale.authKeyFile is set -- nothing in this
repo sets one, so the old flags would never have actually been applied.
extraSetFlags runs unconditionally via tailscaled-set on every boot, so
--advertise-exit-node self-reapplies once the operator has done the
one-time manual `tailscale up` auth.
Verified: all three new targets eval cleanly, nixpkgs-fmt/statix clean,
and a dry-run build of lxc-tailscale-exit-node's tarball resolves its full
closure including tailscaled-set.service.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
virtualisation/proxmox-lxc.nix registers the Nix store DB via a systemd
service, never an activation script -- so neededForUsers sops secrets
(password hashes) and the user-creation step that consumes them never ran
on a real first boot, leaving /etc/shadow stuck with build-time placeholder
entries. boot.postBootCommands looked like the right hook (stage-2-init.sh
does invoke it) but switch-to-configuration behaves unreliably that early,
before systemd itself is up. Fixed with a genuine oneshot systemd service,
gated by ConditionPathExists so it only ever runs once.
Confirmed live via a from-scratch destroy+rebuild+redeploy of the
lxc-nix-cache test container: real password hashes applied automatically,
systemctl is-system-running -> running, zero failed units.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
The real root cause behind the original nix-cache 502, traced all the way
through: modules/installer/host-keys.nix (which NIXOS_HOST_KEYS_DIR=...
--impure actually wires up) is only ever imported by the installer's own
modules/installer/common.nix -- modules/platforms/lxc.nix, which every
real lxc-* host build actually uses, never imported anything like it.
docs/auto-installer.md previously claimed NIXOS_HOST_KEYS_DIR bakes a key
into lxc-* tarballs "the same way it does for the ISO/PXE installer
images" -- that was never actually true; I wrote it without verifying the
mechanism existed for lxc.nix specifically.
In practice this meant every lxc-* container booted with a freshly
self-generated SSH host key that could never match whatever .sops.yaml
actually trusts for that target, so *every* secret -- not just
cache-priv-key -- silently failed to decrypt. No error surfaces in the
boot log for this: the activation step that installs secrets only runs
on a genuinely fresh first activation and silently no-ops once
/run/current-system already exists, so by the time anyone looks the
window has closed. Found by manually invoking sops-install-secrets
directly: "Error getting data key: 0 successful groups required, got 0".
Fixed by giving modules/platforms/lxc.nix the same key-baking mechanism
the installer has, but keyed to its own exact flake target and placing
the key directly at /etc/ssh/ssh_host_ed25519_key (no copy step to stage
for, unlike the installer's /etc/host-keys/ staging area -- an lxc-*
tarball has no install step). The target name comes in via
specialArgs.flakeTarget (new, set by flake.nix's mkTarget) rather than
being read back from config.environment.etc."flake-target" -- reading
that back from within a module that also contributes to
environment.etc is circular (confirmed: "infinite recursion
encountered").
Verified live end-to-end against the real test container (lxc-nix-cache,
VMID 100 on pve.sweet.home): destroyed it, rebuilt the tarball fresh with
the fix, recreated it, and confirmed /run/secrets/ now has all three
secrets this host needs (beszel-token, cache-priv-key, nix-github-token),
nix-serve is active (running), and curl http://localhost/nix-cache-info
succeeds both directly and through nginx.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
exit-node.nix advertised --advertise-exit-node/--advertise-routes but
never enabled IP forwarding, so it wouldn't actually route traffic.
services.tailscale.useRoutingFeatures = "server" is the built-in way to
get the sysctls right without duplicating modules/networking/enable-ip-forwarding.nix.
Also opens the firewall for direct (non-DERP-relayed) peer connections,
and now imports enable-service.nix instead of duplicating `enable = true`,
so "plain tailscale" and "tailscale exit node" are two composable modules.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
nix-serve's secretKeyFile was a manual, undocumented-outside-a-comment
`nix-store --generate-binary-cache-key` step per host -- easy to miss on
a fresh nix-cache instance (as lxc-nix-cache testing just found: systemd
fails the unit with EXIT_CREDENTIALS when LoadCredential can't find the
source file, which nginx then reports as a 502 from clients). It also
can't be regenerated per-host safely: modules/nix-cache/client.nix
hardcodes every client's trust in one specific public key, so every
nix-cache instance has to share the exact same keypair.
Sourced from secrets/nix-cache.yaml's new cache-priv-key entry instead,
via the same sops-nix pattern every other secret in this repo already
uses. Verified the added value derives to the exact public key
modules/nix-cache/client.nix already trusts before committing.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Adds nested vars.nfsShares (subpath + mountpoint per dataset, previously
duplicated independently across server.nix's NFS exports, mount-data.nix's
client mounts, docker.nix's tmpfiles rules, traefik's log rotation path,
and hosts/server/host.nix's beszel config), vars.ports (every literal port
in modules/ and hosts/, kept as separate entries per service even where
numbers coincide so changing one can't silently change another), plus
vars.proxmoxImageSize, vars.nixCacheGcMaxAge, vars.traefikLogRotate, and
raspberryPiHost/raspiNfsPath/raspiMountpoint for the Pi's own NFS export.
Also fixes docker.nix/minimal.nix/gui.nix hardcoding the literal "nixos"
username instead of the existing vars.primaryUser, found during the sweep.
system.stateVersion is deliberately left untouched everywhere -- per
NixOS's own docs that value must stay fixed from first install, not
follow any shared variable.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Linode provisions and sizes /dev/sda (root) and /dev/sdb (swap) itself
as whole, unpartitioned block devices before the OS ever boots.
modules/disko/linode.nix declares them with destroy = false (skips
Disko's wipe stage for these disks entirely) and a bare filesystem/swap
content type matching that existing layout, so re-running it against an
already-provisioned disk only mkfs/mkswaps if blkid shows it isn't
formatted yet -- never repartitions or destroys data.
With every host reachable through the installer menu now carrying a
Disko config, auto-install.sh no longer needs to probe the flake and
branch between `disko --mode destroy,format,mount` and a bind-mount
fallback -- it just always runs Disko.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Both existed only so the installer could boot as an LXC container and
nixos-install some other host from within it, but lxc-* targets are
already excluded from the install menu (nixos-install can't touch its
own running root filesystem), and now have their own direct tarball
path anyway. That left the installer's own LXC form with no real use
case, and packages.all with only two members worth bundling.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
modules/disko/proxmox.nix gains imageSize (20G default) and a
per-host imageName (networking.hostName, so every proxmox-* host
produces a distinctly named image instead of an identical main.raw).
This is the same disko.devices config already used to format a real
disk on install, so it's available for every proxmox-* target with no
per-host changes needed:
nix build .#nixosConfigurations.<host>.config.system.build.diskoImagesScript
sudo ./result --build-memory 2048
docs/proxmox-images.md covers building, host-key pre-seeding via
disko's --pre-format-files (same host-keys/ workflow as the installer
and LXC tarball paths), and the qm import/attach sequence for
deploying the result to Proxmox.
Also fixes a real bug in auto-install.sh found while testing: the
disko confirmation bypass used --yes, which disko's CLI doesn't
recognize at all (the actual flag is --yes-wipe-all-disks) — so the
"skip confirmation" flag was silently a no-op and the interactive
prompt kept appearing regardless.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
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
Finishes the nix-auto-installer migration: scripts/prepare-host-key.sh
and the local host-keys/ directory (gitignored, private key material,
never committed — moved as plain files, not through git history)
weren't carried over in the initial migration.
Also implements automatic key staging, replacing the manual
scp-after-boot step:
- modules/installer/host-keys.nix reads host-keys/ via
builtins.getEnv, which Nix silently returns as "" under normal
(non---impure) evaluation — the module is a no-op by default, safe
for CI, until explicitly opted into:
NIXOS_HOST_KEYS_DIR=$(pwd)/host-keys nix build .#iso --impure
When built this way every key present gets baked into the image at
/etc/host-keys/, and auto-install.sh installs whichever one matches
the flake target selected at install time — no manual per-host scp.
- This deliberately includes the PXE netboot variant, even though
pxe-boot serves it unauthenticated over LAN HTTP: accepted
explicitly as a reasonable trade-off for a network that sits behind
LAN-only infrastructure, not the open internet. auto-install.sh
still falls back to /root/host-keys (manual scp) if a key isn't
baked in, so images built without --impure keep working exactly as
before.
- docs/auto-installer.md replaces nix-auto-installer's README,
updated for in-repo paths and the new build flow.
Verified: normal `nix eval` (no --impure) evaluates identically across
all 19 nixosConfigurations + 4 packages with zero host-keys/* entries
(CI-unaffected); with --impure + the env var set, all three installer
variants (installer/ISO, proxmox-lxc, pxe) correctly embed every key
in host-keys/.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Adds debian@surface's existing SSH key (already used as its admin
key elsewhere in this repo) to vars.remoteBuilderAuthorizedKeys so
nix-cache will accept it as a distributed-build client once deployed.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Zero W20 (repeated attribute keys), W10 (empty { ... }: variadic
pattern, use _: instead), and W04 (a = x.a instead of inherit)
warnings remain anywhere in the tree, and nixpkgs-fmt --check is
clean on all 46 .nix files.
Repeated-key merges go as deep as statix actually flags per file
(e.g. boot.loader.* nested under boot.loader = { ... } once the
outer boot.* merge exposed it as its own repeat) — every merge is a
pure attribute-path restructuring with no value changes, verified by
comparing config.system.build.toplevel.drvPath before/after for a
representative host per changed module plus a full 19-host + 4-package
eval sweep.
One indentation slip caught and fixed during this pass: nesting
modules/installer/common.nix's environment.etc."auto-install.sh".text
under an environment = { ... } block initially normalized the
script's shebang/set line indentation, which actually changes the
rendered file (Nix's '' string dedent treats it as real content, not
cosmetic whitespace) — reproduced the original's exact indentation
and reverified the rendered script is byte-identical to before.
modules/services/zfs/auto-mount-volumes.nix picked up formatting too;
worth noting it isn't imported by anything in this flake at all.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Folds the separate nix-auto-installer repo's build outputs into this
flake so it can build every auto-installer artifact it used to,
negating the need for that repo:
- modules/installer/{common,iso,proxmox-lxc}.nix — migrated from
nix-auto-installer's common.nix/installer.nix/proxmox-lxc.nix.
- flake.nix gains nixosConfigurations.{installer,proxmox-lxc} and
packages.x86_64-linux.{iso,lxc,pxe,all}, matching the original
repo's interface (nix build .#iso / .#lxc / .#pxe / .#all).
- Dropped the live Gitea PAT baked into every installer image via
environment.etc."git-credentials" — gitea.lan.ddnsgeek.com/beatzaplenty/nixos.git
allows anonymous read, so the installer (which only ever reads the
flake, never pushes) doesn't need a credential at all.
- installer_old.nix wasn't migrated — dead code, unreferenced by the
source repo's flake.nix, and carried a second stale leaked token.
- The installer environment's own hardcoded login password hash is
preserved as-is: sops-nix has no stable per-boot host key to derive
an age key from on ephemeral installer media, so it can't reuse the
same per-host secret mechanism the rest of this repo uses without
separate design work.
- vars.adminSshKey / vars.dockerHost-style dedup: the installer's SSH
authorized key and its FLAKE_BASE_URL domain were exact duplicates
of values already in variables.nix / modules/common/configuration.nix,
so both now reference the single source of truth instead.
Verified eval-equivalent for every existing host (drvPath-identical)
and confirmed the migrated auto-install.sh script renders byte-for-byte
identical to the source repo's output.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Several single-purpose modules sat at modules/ root or in the
services/ catch-all despite the repo's established pattern of one
directory per concern (tailscale/, beszel/, docker/, nix-cache/):
- remote-builder-client.nix -> nix-cache/ (always co-included with
nix-cache/client.nix in flake.nix's mkTarget, same buildType guard)
- set-locale.nix -> common/ (unconditionally imported by
common/configuration.nix already)
- enable-ip-forwarding.nix -> networking/
- rotate-traefik-logs.nix -> traefik/rotate-logs.nix
- services/docker-health-to-gotify.nix and services/nextcloud-cron-job.nix
-> docker/ (both only ever imported by the docker build type, same
as the rest of modules/docker/*)
Pure path moves plus import-path updates in flake.nix,
common/configuration.nix, and build-types/docker.nix — verified
eval-equivalent (drvPath-identical) across representative hosts.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
hosts/server/host.nix and hosts/nix-cache/host.nix each hand-rolled the
same sops secret/template/environmentFile wiring for the beszel agent
token, differing only in the sops file path and template name. Factor
it into modules/beszel/host-token.nix ({ name, sopsFile }) so a third
host can adopt it without copy-pasting the boilerplate again. Also
drops two dead, stale commented-out HUB_URL lines left over from
before variables.nix grew a homeDomain var.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
Pulls the beszel hub / PVE / PBS ports, the docker-compose host's LAN
name, and the remote-builder client SSH keys out of scattered inline
literals across modules/hosts and into variables.nix as the single
source of truth, matching the existing pattern for other cross-host
references (nixCacheHost, nfsServerHost).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
boot.isContainer disables services.udev, which NetworkManager depends on
to enumerate devices — this left NM unable to reliably manage the
container veth, breaking DHCP-hostname registration in Pi-hole. It also
defaulted networking.useHostResolvConf to true, which assumes a
systemd-nspawn-style resolv.conf bind-mount that real Proxmox LXC doesn't
provide (nixpkgs' own proxmox-lxc.nix module forces this false for the
same reason). Also guard the networkmanager extraGroups membership in the
minimal/gui build-types, since that group only exists when NM is enabled.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01La55Nsss8jZ7ZuzUV9mfot
LXC containers share the host kernel — Proxmox starts them by exec'ing
/sbin/init directly in the container's rootfs, no bootloader or initrd
involved at all. Without boot.loader.initScript.enable, that file
isn't wired to launch the current generation, so even a correctly
installed system (see the nix-auto-installer bind-mount fix, same
underlying issue) could still fail to come up after reboot. This is
exactly what nixpkgs' own virtualisation/proxmox-lxc.nix module sets
for the same reason.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
One file (variables.nix) holding every value that was previously
hardcoded and repeated across modules: LAN domain/CIDR, home/tailnet
domains, cross-host references (nix-cache substituter hostname, NFS
server hostname, remote-builder user), PXE/PBS IPs, timezone, and the
primary username.
Wired in via flake.nix's specialArgs (and home-manager's
extraSpecialArgs for the two home.nix files), so any module picks it
up by just adding `vars` to its function arguments — no explicit
import needed. Two hosts (nix-cache, server) now derive their own
networking.hostName from the same variable other hosts use to reach
them, so there's exactly one place to change either identifier.
Purely mechanical: every substituted value matches what was already
there, confirmed by identical toplevel .drv paths for all 17 targets
before and after.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Audited the working tree and full git history for committed secrets
(gitleaks + trufflehog + manual grep, see secrets-inventory.md, kept
local/gitignored per the spec). Found: a password hash shared by root
and the nixos user across every host, two live Beszel monitoring
tokens, and a GitHub fine-grained PAT embedded in a home-manager
nix.conf.
Migrates all of them to sops-nix:
- .sops.yaml + secrets/*.yaml, encrypted for admin + the age keys
derived (via ssh-to-age) from each live host's existing SSH host
key — no new key material transferred to any machine.
- users.users.{root,nixos}.hashedPasswordFile replaces the inline
hashedPassword shared by every target.
- The GitHub PAT moves from a home-manager-managed, store-visible
nix.conf to a sops.templates-rendered file included via nix.conf's
native !include, system-wide instead of per-user.
- Beszel TOKEN moves from `environment` (store-visible) to
`environmentFile` (runtime-only via sops.templates); the dead
commented-out docker token is removed from the tree entirely.
Added a tracked pre-commit hook (gitleaks protect --staged, wired via
core.hooksPath) so a secret can't be committed by accident again, and
documented the sops workflow in README.md.
Structural verification only: all 17 flake targets evaluate, and
`nix build --dry-run --no-link` succeeds for the three currently
deployed hosts. Per CLAUDE.md, actual `nixos-rebuild switch` — the
step that confirms secrets decrypt and services start on a real
machine — is left for manual verification.
Git history still contains the original plaintext secrets; scrubbing
history (Milestone 3) and rotating every credential (Milestone 4) are
separate, deliberately gated steps per remove-sensetive-info-refactor.md.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Generates all nixosConfigurations from mkTarget(platform, buildType,
hostPath) instead of hand-written per-host blocks, so adding a new
platform or build type is a one-line addition. Per-machine identity
(hostname, hostId, secrets, stateVersion) moves into hosts/<name>/host.nix;
platform-specific config (hardware, boot, networking) into
modules/platforms/*.nix; build-type config (minimal/server/docker/gui/
nix-cache/pxe-boot) into modules/build-types/*.nix.
Old flat targets (nixos, docker, server, nix-cache, nix-minimal, pxe-boot)
are replaced by the 17-target <platform>-<buildtype> matrix; each new
target was verified to evaluate before its old counterpart was removed.
CI workflows and docs/aliases now discover hosts dynamically via
nixosConfigurations attrNames and /etc/flake-target instead of hardcoded
lists, so they can't drift from flake.nix again.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
beatzaplentyandClaude Sonnet 5