nixpkgs/nixos/modules/services/networking/tailscale.nix

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

104 lines
3.7 KiB
Nix
Raw Normal View History

{ config, lib, pkgs, ... }:
with lib;
let
cfg = config.services.tailscale;
isNetworkd = config.networking.useNetworkd;
in {
meta.maintainers = with maintainers; [ danderson mbaillie twitchyliquid64 ];
options.services.tailscale = {
enable = mkEnableOption (lib.mdDoc "Tailscale client daemon");
port = mkOption {
type = types.port;
default = 41641;
description = lib.mdDoc "The port to listen on for tunnel traffic (0=autoselect).";
};
interfaceName = mkOption {
type = types.str;
default = "tailscale0";
description = lib.mdDoc ''The interface name for tunnel traffic. Use "userspace-networking" (beta) to not use TUN.'';
};
permitCertUid = mkOption {
type = types.nullOr types.nonEmptyStr;
default = null;
description = lib.mdDoc "Username or user ID of the user allowed to to fetch Tailscale TLS certificates for the node.";
};
package = mkOption {
type = types.package;
default = pkgs.tailscale;
defaultText = literalExpression "pkgs.tailscale";
description = lib.mdDoc "The package to use for tailscale";
};
useRoutingFeatures = mkOption {
type = types.enum [ "none" "client" "server" "both" ];
default = "none";
example = "server";
description = lib.mdDoc ''
Enables settings required for Tailscale's routing features like subnet routers and exit nodes.
To use these these features, you will still need to call `sudo tailscale up` with the relevant flags like `--advertise-exit-node` and `--exit-node`.
When set to `client` or `both`, reverse path filtering will be set to loose instead of strict.
When set to `server` or `both`, IP forwarding will be enabled.
'';
};
};
config = mkIf cfg.enable {
environment.systemPackages = [ cfg.package ]; # for the CLI
systemd.packages = [ cfg.package ];
systemd.services.tailscaled = {
wantedBy = [ "multi-user.target" ];
path = [
config.networking.resolvconf.package # for configuring DNS in some configs
pkgs.procps # for collecting running services (opt-in feature)
pkgs.glibc # for `getent` to look up user shells
];
serviceConfig.Environment = [
"PORT=${toString cfg.port}"
''"FLAGS=--tun ${lib.escapeShellArg cfg.interfaceName}"''
] ++ (lib.optionals (cfg.permitCertUid != null) [
"TS_PERMIT_CERT_UID=${cfg.permitCertUid}"
]);
# Restart tailscaled with a single `systemctl restart` at the
# end of activation, rather than a `stop` followed by a later
# `start`. Activation over Tailscale can hang for tens of
# seconds in the stop+start setup, if the activation script has
# a significant delay between the stop and start phases
# (e.g. script blocked on another unit with a slow shutdown).
#
# Tailscale is aware of the correctness tradeoff involved, and
# already makes its upstream systemd unit robust against unit
# version mismatches on restart for compatibility with other
# linux distros.
stopIfChanged = false;
};
boot.kernel.sysctl = mkIf (cfg.useRoutingFeatures == "server" || cfg.useRoutingFeatures == "both") {
"net.ipv4.conf.all.forwarding" = mkDefault true;
"net.ipv6.conf.all.forwarding" = mkDefault true;
};
networking.firewall.checkReversePath = mkIf (cfg.useRoutingFeatures == "client" || cfg.useRoutingFeatures == "both") "loose";
networking.dhcpcd.denyInterfaces = [ cfg.interfaceName ];
systemd.network.networks."50-tailscale" = mkIf isNetworkd {
matchConfig = {
Name = cfg.interfaceName;
};
linkConfig = {
Unmanaged = true;
ActivationPolicy = "manual";
};
};
};
}