Page not found, maybe moved. Try searching or go to homepage.
+ +Hi. I’m Edgar, more often known as Guekka, G’k or G_ka. I like Nix(OS), Rust, C++, and a lot of other technologies.
+My most important contribution to the software world is an automation tool called Cathedral Assets Optimizer (source) which is used by thousands of people. Its main purpose is helping the process of “porting” a mod between Skyrim Legendary Edition and Skyrim Special Edition.
+ ++ In the previous part, we configured our NixOS server to use impermanence. I have made a few changes since, most notably moving to a proper VM in Proxm… + ...more +
+ A few months ago, I woke up with the idea of hosting my own services. I went through a lot of tries. LXC, Debian, Alpine, (rootless or not) Docker, po… + ...more +
').concat((0,r.highlight)(e.entry.title,e.title_highlight_ranges||[]),"
\n ").concat(t.showScores?"".concat(e.score,""):"","\n \n ...'.concat((0,r.highlight)(e.text,e.highlight_ranges||[]),"...\n
\n ").concat(t.showScores?"".concat(e.score,""):"","\n ').concat((0,r.highlight)(e.entry.title,e.title_highlight_ranges||[]),"
\n ").concat(t.showScores?"".concat(e.score,""):"","\n \n ...'.concat((0,r.highlight)(e.text,e.highlight_ranges||[]),"...\n
\n ").concat(t.showScores?"".concat(e.score,""):"","\n A few months ago, I woke up with the idea of hosting my own services. I went through a lot of tries. LXC, Debian, Alpine, (rootless or not) Docker, podman, portainer…
+But no solution felt perfect. Until I decided to have a try at hosting using NixOS.
+I’m going to assume you know about NixOS and have some prior experience. However, for a small summary: NixOS is a Linux distribution revolving around the Nix package manager. Its main advantage is having a reproducible environment through a declarative configuration. This means that you can copy an entire computer configuration easily: if it works somewhere, it will work anywhere.
+My main focus point is reproducibility, so that’s why we’ll start with configuring impermanence.
+Originally, a philosophic concept. But in our case, impermanence means erasing the / drive at each reboot. You read that right, erasing almost everything at each reboot. This part stands on the shoulders of those who did it before me:
The goal is the following: over years, configuration files accumulate. Sometimes editing /etc is required, because of a bug or an obscure configuration. NixOS allows us to avoid this manual file editing, but it does not force us to do so. We can still have a lot of important state, breaking the reproducibility promise.
So what can we do instead? Erase everything, at each reboot. This way, we’ll be sure the only source of truth is our configuration.
+I’m currently using a quickemu VM. This is not a recommenced setup and is only done for testing. Configuration file:
+"linux"
+"nixos-22.11-minimal/disk.qcow2"
+"nixos-22.11-minimal/latest-nixos-minimal-x86_64-linux.iso"
+"50G"
+"4G"
+Let’s first format it:
+DISK=/dev/vda
+
+
+
+
+
+
+
+
+++Using swap in 2023!?
+
Yes.
+
+
+While the impermanence module recommends using tmpfs for /, I chose to use btrfs: I do not have RAM to waste. Furthermore, this will allow us to use a nice script we’ll see later on.
Let’s create btrfs subvolumes:
+
+
+
+
+
+And now, the crucial part:
+
+We just took a snapshot of that empty volume. We will restore it at each reboot.
+We can now mount the subvolumes and let nixos-generate-config do its job
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+Lastly, we only have to edit the generated configuration files at /mnt/etc/nixos.
My final configuration is available here. You can follow all the steps by looking at the commits.
+/mnt/etc/nixos/hardware-configuration.nix.I’ve added "compress=zstd" "noatime" to all filesystems. We also need to add neededForBoot to /var/log and /persist.
configuration.nixI’ve enabled networkmanager, removed most suggested options and enabled system.copySystemConfiguration.
This last option copies the current configuration to /run/current_system/configuration.nix. You should not rely on it: keep your configuration in a git repository. But it can serve as some kind of last chance.
users.mutableUsers = false;
+users.users.user = {
+ isNormalUser = true;
+ extraGroups = [ "wheel" ];
+
+ openssh.authorizedKeys.keys = [ "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAICWVNch9BcjkMqS/Xwep+GN4HwqyRIjr3Cuw7mHpqsKr nixos" ];
+
+ # passwordFile needs to be in a volume marked with `neededForBoot = true`
+ passwordFile = "/persist/passwords/user";
+};
+Here we have completely disabled imperative user modification. This does not matter much, as imperative changes would be erased anyway at start.
+We thus need to provide a password. We’re using passwordFile for that: a path to a file containing the hashed password.
Here’s how to generate that file: sudo mkpasswd -m sha-512 "hunter2" > /mnt/persist/passwords/user.
The SSH key was generated using `ssh-keygen -t ed25519 -C “nixos”.
+ services.openssh = {
+ enable = true;
+ passwordAuthentication = false;
+ allowSFTP = false; # Don't set this if you need sftp
+ challengeResponseAuthentication = false;
+ extraConfig = ''
+ AllowTcpForwarding yes
+ X11Forwarding no
+ AllowAgentForwarding no
+ AllowStreamLocalForwarding no
+ AuthenticationMethods publickey
+ '';
+ };
+This reduces attack surface, for example by disabling stream-local forwarding and disabling password authentification.
+This will be enough for now. Let’s install the system before going to the next step: sudo nixos-install --root /mnt && sudo reboot. You should be able to connect by SSH using the previously defined key, or login using the password you defined in /persist/passwords/user.
We’ve created our volumes, we’ve configured the system… But I promised we would reset our system at each reboot. Let’s do that now!
+We’re going to use the following script, credit of mt-caret. Do not forget to replace vda3 with your data partition.
16/07/23 update: it was brought to my attention that postDeviceCommands can cause data loss. +While I did not experience any issue, I have updated the script to use a safer alternative.
+ boot.initrd = {
+ enable = true;
+ supportedFilesystems = [ "btrfs" ];
+
+ systemd.services.restore-root = {
+ description = "Rollback btrfs rootfs";
+ wantedBy = [ "initrd.target" ];
+ requires = [
+ "dev-vda3"
+ ];
+ after = [
+ "dev-vda3"
+ # for luks
+ "systemd-cryptsetup@config.networking.hostName.service"
+ ];
+ before = [ "sysroot.mount" ];
+ unitConfig.DefaultDependencies = "no";
+ serviceConfig.Type = "oneshot";
+ script = ''
+ mkdir -p /mnt
+
+ # We first mount the btrfs root to /mnt
+ # so we can manipulate btrfs subvolumes.
+ mount -o subvol=/ /dev/vda3 /mnt
+
+ # While we're tempted to just delete /root and create
+ # a new snapshot from /root-blank, /root is already
+ # populated at this point with a number of subvolumes,
+ # which makes `btrfs subvolume delete` fail.
+ # So, we remove them first.
+ #
+ # /root contains subvolumes:
+ # - /root/var/lib/portables
+ # - /root/var/lib/machines
+ #
+ # I suspect these are related to systemd-nspawn, but
+ # since I don't use it I'm not 100% sure.
+ # Anyhow, deleting these subvolumes hasn't resulted
+ # in any issues so far, except for fairly
+ # benign-looking errors from systemd-tmpfiles.
+ btrfs subvolume list -o /mnt/root |
+ cut -f9 -d' ' |
+ while read subvolume; do
+ echo "deleting /$subvolume subvolume..."
+ btrfs subvolume delete "/mnt/$subvolume"
+ done &&
+ echo "deleting /root subvolume..." &&
+ btrfs subvolume delete /mnt/root
+
+ echo "restoring blank /root subvolume..."
+ btrfs subvolume snapshot /mnt/root-blank /mnt/root
+
+ # Once we're done rolling back to a blank snapshot,
+ # we can unmount /mnt and continue on the boot process.
+ umount /mnt
+ '';
+ };
+ };
+We can then specify the files we want to keep.
+But which files do we want to keep? Let’s find out. Thanks to another useful script of mt-caret, we can list the differences between our current / and the blank state:
#!/usr/bin/env bash
+# fs-diff.sh
+
+
+OLD_TRANSID=
+OLD_TRANSID=
+
+ |
+ |
+ |
+ |
+ |
+while ; do
+ path="/"
+ if [; then
+ # The path is a symbolic link, so is probably handled by NixOS already
+ elif [; then
+ # The path is a directory, ignore
+ else
+
+ fi
+done
+Used like this:
+ ; ;
+Here’s the result of mine:
+/etc/.clean
+/etc/group
+/etc/machine-id
+/etc/nixos/configuration.nix
+/etc/nixos/hardware-configuration.nix
+/etc/passwd
+/etc/resolv.conf
+/etc/shadow
+/etc/ssh/authorized_keys.d/user
+/etc/ssh/ssh_host_ed25519_key
+/etc/ssh/ssh_host_ed25519_key.pub
+/etc/ssh/ssh_host_rsa_key
+/etc/ssh/ssh_host_rsa_key.pub
+/etc/subgid
+/etc/subuid
+/etc/sudoers
+/etc/.updated
+/root/.nix-channels
+/root/.nix-defexpr/channels
+/var/lib/NetworkManager/internal-84e273c2-b91a-3a96-b341-8234a339bdc7-enp0s8.lease
+/var/lib/NetworkManager/internal-84e273c2-b91a-3a96-b341-8234a339bdc7-enp0s9.lease
+/var/lib/NetworkManager/NetworkManager-intern.conf
+/var/lib/NetworkManager/secret_key
+/var/lib/NetworkManager/timestamps
+/var/lib/nixos/auto-subuid-map
+/var/lib/nixos/declarative-groups
+/var/lib/nixos/declarative-users
+/var/lib/nixos/gid-map
+/var/lib/nixos/uid-map
+/var/lib/systemd/catalog/database
+/var/lib/systemd/random-seed
+/var/.updated
+That’s not too bad!
+Out of these, there’s almost nothing I want to preserve.
+Let’s make use of the impermanence module. We need to download it:
let
+ impermanence = builtins.fetchTarball "https://github.com/nix-community/impermanence/archive/master.tar.gz";
+in
+{
+imports = [ "impermanence/nixos.nix" ./hardware-configuration.nix ]
+// the whole configuration
+}
+And now, we can just tell it the files and directories that we want:
+ # configure impermanence
+ environment.persistence."/persist" = {
+ directories = [
+ "/etc/nixos"
+ ];
+ files = [
+ "/etc/machine-id"
+ "/etc/ssh/ssh_host_ed25519_key"
+ "/etc/ssh/ssh_host_ed25519_key.pub"
+ "/etc/ssh/ssh_host_rsa_key"
+ "/etc/ssh/ssh_host_rsa_key.pub"
+ };
+
+ security.sudo.extraConfig = ''
+ # rollback results in sudo lectures after each reboot
+ Defaults lecture = never
+ '';
+What an ergonomic interface.
+++Wait, did you just say Nix was ergonomic?
+
Well, yes. Sometimes.
+I have not saved my network manager configuration, but you may need to.
+When new files are set to be preserved, it is necessary to copy them manually to /persist:
+
+
+
+
+
+
+
+
+
+
+
+
+Now, if we reboot and list files again:
+/etc/.clean
+/etc/group
+/etc/passwd
+/etc/resolv.conf
+/etc/shadow
+/etc/ssh/authorized_keys.d/user
+/etc/subgid
+/etc/subuid
+/etc/sudoers
+/etc/.updated
+/root/.nix-channels
+/var/lib/NetworkManager/internal-84e273c2-b91a-3a96-b341-8234a339bdc7-enp0s9.lease
+/var/lib/NetworkManager/NetworkManager-intern.conf
+/var/lib/NetworkManager/secret_key
+/var/lib/NetworkManager/timestamps
+/var/lib/nixos/auto-subuid-map
+/var/lib/nixos/declarative-groups
+/var/lib/nixos/declarative-users
+/var/lib/nixos/gid-map
+/var/lib/nixos/uid-map
+/var/lib/systemd/catalog/database
+/var/lib/systemd/random-seed
+/var/.updated
+Success! The files we persisted are no longer showing up.
+It is possible to setup the impermanence module for our home directory. However, I did not want to go through home-manager installation. Furthermore, a home directory is meant to be stateful.
In our case, we are creating a server, so it would still make sense to configure it. If you are interested, have a look at tmpfs at home.
+In the next part, we will make our server more secure by making it only available through Tailscale. We will also setup our first service.
+I hope you’ve enjoyed this article! Thanks for reading to the end!
+ + + + + + + + + + +In the previous part, we configured our NixOS server to use impermanence. I have made a few changes since, most notably moving to a proper VM in Proxmox.
+The following instructions might lack some details, but you can follow the GitHub repo to see the full code.
+++If you already know about flakes, you can safely ignore this part.
+
Have you heard about Nix flakes? If you have been in the Nix ecosystem for more than a few days, most likely. They’re the shiny new way to write Nix code, still experimental but used everywhere.
+Their main advantage over traditional Nix is purity, mainly with their defined inputs and outputs.
Remember when I told you Nix was reproducible? It was a lie. Let me explain myself: when writing Nix code, we always have some kind of input. For example, nixpkgs will be required almost all the time. There are two ways to obtain it.
import builtins.fetchTarball "https://github.com/nixos/nixpkgs/archive/nixos-22.11.tar.gz";import <nixpkgs> {}This second way uses a globally-defined configuration, which can change externally to our Nix files. We thus lose complete reproducibility. Instead, flakes allow us to avoid channels by specifying inputs alongside Nix configuration, as well as blocking some actions that could hinder reproducibility.
+For a more in depth introduction, have a look at the wiki.
+Now, why do we want to migrate to flakes? We do not have external requirements, do we?
+Well, yes, we do. Apart from the obvious nixpkgs dependency, which is configured system-wide, the impermanence module is being imported:
impermanence = builtins.fetchTarball "https://github.com/Nix-community/impermanence/archive/master.tar.gz";
+This fetchTarball call is bad, by the way, as we do not specify the expected hash. We could be a victim of a man-in-the-middle attack and not notice it.
Now that I intend to add more modules, and possibly use the unstable channel, it is better to migrate. Let’s see what our entry point would look like:
{
+ # what is consumed (previously provided by channels and fetchTarball)
+ inputs = {
+ nixpkgs.url = "github:NixOS/nixpkgs/nixos-22.11"; # (1)
+ impermanence.url = "github:Nix-community/impermanence";
+ };
+
+ # what will be produced (i.e. the build)
+ outputs = { nixpkgs, ... }@inputs: { # (2)
+ nixosConfigurations = { # (3)
+ server = nixpkgs.lib.nixosSystem { # (4)
+ packages = nixpkgs.legacyPackages.x86_64-linux; # (5)
+ specialArgs = inputs; # forward inputs to modules
+ modules = [
+ ./configuration.Nix
+ ];
+ };
+ };
+ };
+}
+That’s a lot to understand at once. Let’s study it one line at a time.
+Firstly, we have to understand this file is simply describing outputs as a function taking inputs. Like in mathematics, we create the same output given the same input: a pure function, would say functional programmers.
(1) is defining an input: we simply give it an url. That line can be translated as *Grab the nixos-22.11 branch from the GitHub repo nixpkgs owned by NixOS.
(2) is defining the outputs function. Most complicated things are defined using functions in Nix. It takes a named attribute set argument as an input. So this syntax:
{ nixpkgs, ... }@inputs: nixpkgs
+is the same as:
+inputs: inputs.nixpkgs
+In both cases, we’re accessing the nixpkgs property of the inputs set.
But in the first case, we don’t have to repeat inputs. everywhere. In JS, you would call that destructuring: it is just making inner elements easier to access. If you have troubles understanding the Nix syntax, I personally like FasterThanLime article.
(3): NixOS configuration have to be placed specifically in the nixosConfigurations set.
(4) is the place where we actually define the system. We call the nixosSystem function and pass it some arguments. Yes, the whole system is an output too!
(5): we give the packages instance to our system. In our case, we are passing the default packages, but we might want to modify them before. We also have to specify our architecture (x86_64).
That’s pretty much it. With that flake.Nix in /etc/nixos, nixos-rebuild will work as before. However, if you’re using git, beware that your files all need to be under version control or Nix will not see them.
In order to set up Tailscale, we will use a pre-auth key. This allows us to connect to our server without interaction. However, we must hide this key, or other people could join our Tailscale network, which could obviously have dangerous consequences.
+There are 2 well-known solutions : agenix and sops-nix. I’ve chosen sops for no particular reason. +The first step will be to add it to our flake. See, we already get a use for it!
+sops-nixLet’s change our inputs:
inputs = {
+ # ...
+ sops-nix = {
+ url = "github:mic92/sops-nix";
+ inputs.nixpkgs.follows ="nixpkgs";
+ };
+ };
+FollowsWhat’s up with this follows? sops-nix already depends on nixpkgs, but it might use a different revision than ours. Making it use our own has several advantages:
And how do we know if a package has inputs that need to be redirected? That’s the neat thing, we don’t. Either we have to look at the upstream flake.nix, or we can call nix flake info and get a graph like so:
Resolved URL: git+file:///etc/nixos
+Locked URL: git+file:///etc/nixos
+Path: /Nix/store/4b14z6ki7av3kid69sp5vgf50wzd3a73-source
+Last modified: 2023-04-17 14:04:13
+Inputs:
+├───impermanence: github:Nix-community/impermanence/6138e
+├───nixpkgs: github:NixOS/nixpkgs/39fa0
+└───sops-Nix: github:mic92/sops-nix/de651
+ ├───nixpkgs follows input 'nixpkgs'
+ └───nixpkgs-stable: github:NixOS/nixpkgs/1040c
+We can notice sops-nix also has a nixpkgs-stable input, that we might as well redirect.
sops-nix works by encrypting our secrets with private keys.
+We thus need to provide it with the keys we will use. We can generate an age key, or get one from our SSH host key.
+Each secrets group can have different allowed keys, so that one user cannot access another’s secrets.
+I will use the SSH host key for my server:
$ nix-shell -p ssh-to-age --run 'cat /etc/ssh/ssh_host_ed25519_key.pub | ssh-to-age'
+age1dt24qetqhy2ng53fyj69yq9hg8rdsg4ep0lvvhdg69xw9v4l0asqj6xzkh
+We now have to write .sops.yaml file in order to configure which keys can access which secrets.
keys:
+ - age1dt24qetqhy2ng53fyj69yq9hg8rdsg4ep0lvvhdg69xw9v4l0asqj6xzkh
+creation_rules:
+ - path_regex: hosts/horus/secrets.yaml$
+ key_groups:
+ - age:
+ - *horus
+ - path_regex: hosts/common/secrets.yaml$
+ key_groups:
+ - age:
+ - *horus
+That’s it for decryption. However, we need to write secrets too. For that, we can get the corresponding private key:
+
+That install bit is here to create the directory if it doesn’t exist and set the right permissions.
++Isn’t this insecure? The key is not password-locked.
+
Indeed, if someone has access to our user account, they can read that key and decrypt the secrets. However, we can probably assume our user already has access to the local secrets, so it doesn’t matter much. Our goal is to be able to put these secrets on a public GitHub, not to protect them locally.
+sops-nixOur last step is to configure sops.
+We’re going to get fancy here, as I’m stealing borrowing a module from Misterio’s config. In the future, this will often happen, as his config happens to be a great resource. Let’s have a look at sops.nix:
{ sops-nix, lib, config, ... }:
+let
+ isEd25519 = k: k.type == "ed25519";
+ getKeyPath = k: k.path;
+ keys = builtins.filter isEd25519 config.services.openssh.hostKeys;
+in
+{
+ imports = [
+ sops-nix.nixosModules.sops
+ ];
+
+ sops = {
+ age.sshKeyPaths = map getKeyPath keys;
+ };
+}
+This looks complicated, but it is not. First, we are declaring some functions in the let block.
isEd25519 simply tells if an SSH key uses ed25519getKeyPath gets the path of an SSH keykeys is the list of ed25519 keys, taken from opensshThen we import sops. Finally, we give it the keys we collected earlier. This avoids hardcoding keys, which is great!
We can now import this module in our config:
+ imports =
+ [
+ impermanence.nixosModule
+ ./hardware-configuration.Nix
+ ../../modules/sops.nix
+ ];
+sops-nix is now ready to use. Do not forget to rebuild the config.
Let’s write a secret:
+
+
+An editor should open. We can now write secrets, using yaml. Once we’re done, we can save the file. Example content:
+tailscale_key: e2b6595884993e001da58d2995af65df489582a702e3a2f3
+We now have to tell sops this secret exists. So we declare it somewhere in our configuration:
sops.secrets.tailscale_key = {
+ sopsFile = ./secrets.yaml;
+};
+And that’s all! To use it, we simply have to use config.sops.secrets.tailscale_key.path where we need it. Beware that this will not give you the secret, but a path to a file containing the raw secret, for security reasons. Otherwise, the secret would be in the Nix store, and thus accessible to any user on the system.
If you ever need to add a new host, you will need to update your secrets with sops updatekeys your_secret. This command has to be on a system with already authorized keys.
We can finally get to a real feature, setting up Tailscale. +For those of you who haven’t heard of it, Tailscale is a private meshed network, allowing you to connect to your machines privately and securely through Wireguard, a VPN protocol, without exposing them to the world. +This means being able to close port 22, while still being able to SSH into your computer.
+Furthermore, Tailscale offers some additional features, such as a fancy file sending tool or hole punching, which allows you to connect to your computer even if it is behind a NAT. I won’t go into details here, but you can read more about it on their website.
+I’ve chosen to write a full-fledged NixOS module for Tailscale, as it is a service that needs to be configured and started. This is a good example of a module that can be reused in other configurations, so it’s worth writing it. Let’s get started!
+We’re going to write a module, so we need to create a directory for it:
+
+Inside, we’ll need a default.nix file:
{
+ tailscale-autoconnect = import ./tailscale-autoconnect.nix;
+}
+By convention, this file is automatically imported when you import a directory. Then, in our flake.nix, we can import our module:
outputs = # ...
+{
+ nixosModules = import ./modules/nixos;
+}
+As before, the nixosModules attribute has a special meaning.
Finally, we have to import the module we’re writing in our configuration:
+ imports =
+ [
+ # ...
+ outputs.nixosModules.tailscale-autoconnect
+ ];
+We’re going to write a module to start Tailscale and connect to it automatically. This is a good example of a module that can be reused in other configurations, so it’s worth writing it. Let’s get started!
+First, we need to create a tailscale-autoconnect.nix file in our modules/nixos directory. We’ll start with the boilerplate:
{ config, lib, pkgs, ... }:
+{
+ with lib; let
+ cfg = config.services.tailscaleAutoconnect;
+ in {
+ options = {
+ services.tailscaleAutoconnect = {
+ enable = mkEnableOption "tailscaleAutoconnect";
+ };
+ };
+
+ config = mkIf cfg.services.tailscaleAutoconnect.enable {
+ # ...
+ };
+ };
+}
+This is the basic structure of a module. We declare an option, and then we use it to conditionally change the configuration. So:
+options is the option declarationconfig is the consequence of the option being enabled, the configuration changeLet’s declare all the options first.
+ options.services.tailscaleAutoconnect = {
+ enable = mkEnableOption "tailscaleAutoconnect";
+ authkeyFile = mkOption {
+ type = types.str;
+ description = "The authkey to use for authentication with Tailscale";
+ };
+
+ loginServer = mkOption {
+ type = types.str;
+ default = "";
+ description = "The login server to use for authentication with Tailscale";
+ };
+
+ advertiseExitNode = mkOption {
+ type = types.bool;
+ default = false;
+ description = "Whether to advertise this node as an exit node";
+ };
+
+ exitNode = mkOption {
+ type = types.str;
+ default = "";
+ description = "The exit node to use for this node";
+ };
+
+ exitNodeAllowLanAccess = mkOption {
+ type = types.bool;
+ default = false;
+ description = "Whether to allow LAN access to this node";
+ };
+ };
+This looks like a lot of code, but we’re simply declaring options. We need to give them a type, and we can also give a default value and a description. +Now, the actually useful code:
+ config = mkIf cfg.enable {
+ assertions = [
+ {
+ assertion = cfg.authkeyFile != "";
+ message = "authkeyFile must be set";
+ }
+ {
+ assertion = cfg.exitNodeAllowLanAccess -> cfg.exitNode != "";
+ message = "exitNodeAllowLanAccess must be false if exitNode is not set";
+ }
+ {
+ assertion = cfg.advertiseExitNode -> cfg.exitNode == "";d
+ message = "advertiseExitNode must be false if exitNode is set";
+ }
+ ];
+
+ systemd.services.tailscale-autoconnect = {
+ description = "Automatic connection to Tailscale";
+
+ # make sure tailscale is running before trying to connect to tailscale
+ after = ["network-pre.target" "tailscale.service"];
+ wants = ["network-pre.target" "tailscale.service"];
+ wantedBy = ["multi-user.target"];
+
+ serviceConfig.Type = "oneshot";
+
+ script = with pkgs; ''
+ # wait for tailscaled to settle
+ sleep 2
+
+ # check if we are already authenticated to tailscale
+ status="$(tailscale/bin/tailscale status -json | jq/bin/jq -r .BackendState)"
+ # if status is not null, then we are already authenticated
+ echo "tailscale status: $status"
+ if [ "$status" != "NeedsLogin" ]; then
+ exit 0
+ fi
+
+ # otherwise authenticate with tailscale
+ # timeout after 10 seconds to avoid hanging the boot process
+ coreutils/bin/timeout 10 tailscale/bin/tailscale up \
+ lib.optionalString (cfg.loginServer != "") "--login-server=cfg.loginServer" \
+ --authkey=$(cat "cfg.authkeyFile")
+
+ # we have to proceed in two steps because some options are only available
+ # after authentication
+ coreutils/bin/timeout 10 tailscale/bin/tailscale up \
+ lib.optionalString (cfg.loginServer != "") "--login-server=cfg.loginServer" \
+ lib.optionalString (cfg.advertiseExitNode) "--advertise-exit-node" \
+ lib.optionalString (cfg.exitNode != "") "--exit-node=cfg.exitNode" \
+ lib.optionalString (cfg.exitNodeAllowLanAccess) "--exit-node-allow-lan-access"
+ '';
+ };
+
+ networking.firewall = {
+ trustedInterfaces = [ "tailscale0" ];
+ allowedUDPPorts = [ config.services.tailscale.port ];
+ };
+
+ services.tailscale = {
+ enable = true;
+ useRoutingFeatures = if cfg.advertiseExitNode then "server" else "client";
+ };
+ };
+First, the assertions. They’re here to make sure that the user doesn’t make any mistake when configuring the module. For example, a user cannot both advertise an exit node and set an exit node.
+Then, the service. We’re using systemd to run a script that will connect to Tailscale. The after, wants and wantedBy options make the script run after the network is up and after Tailscale daemon is started. The Type option is here to make sure that the script is run only once. The script itself is a bit long, but it’s just a bunch of bash commands. It’s pretty straightforward. First, we wait for the Tailscale daemon to settle. Then, we check if we’re already authenticated. If we are, we exit. Otherwise, we authenticate. Finally, we connect to Tailscale. We have to do it in two steps because some options are only available after authentication.
At the end, we configure the firewall to allow Tailscale traffic, and we enable the Tailscale service.
+Now, an example of how to use this module:
+{ outputs, ...}:
+{
+ imports = [
+ outputs.nixosModules.tailscale-autoconnect
+ ];
+
+ services.tailscaleAutoconnect = {
+ enable = true;
+ authkeyFile = config.sops.secrets.tailscale_key.path;
+ loginServer = "https://login.tailscale.com";
+ exitNode = "some-node-id";
+ exitNodeAllowLanAccess = true;
+ };
+
+ sops.secrets.tailscale_key = {
+ sopsFile = ./secrets.yaml;
+ };
+
+ environment.persistence = {
+ "/persist".directories = ["/var/lib/tailscale"];
+ };
+}
+The module is imported and configured. We also use the sops secret we created earlier. Finally, we persist the Tailscale state, so that we don’t have to authenticate again after a reboot. This is especially important if the authkey can expire.
That’s all for this post. Thanks for reading! If you have any question, feel free to ask in the comments. The final code can be found here.
+ + + + + + + + + + + +Click here to be redirected.
diff --git a/posts/index.html b/posts/index.html new file mode 100644 index 0000000..38fce8a --- /dev/null +++ b/posts/index.html @@ -0,0 +1,98 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + +Click here to be redirected.
diff --git a/privacy/index.html b/privacy/index.html new file mode 100644 index 0000000..b8ccea4 --- /dev/null +++ b/privacy/index.html @@ -0,0 +1,98 @@ + + + + + + + + + + + + + + + + + + + + + + + + + + +Effective Date: 1st Jan 2022
+ +#nix2 #self-hosting2
+NixOS as a server, part 2: Flake, tailscale -
+NixOS as a server, part 1: Impermanence -
+