nixpkgs-suyu/modules/services/networking/ntpd.nix
Eelco Dolstra 674d63e212 * Don't do a separate "ntp -q -g" to reset the system time in the ntp
start script.  It's probably not necessary, but more importantly, it
  can hang forever if the network is down.  (Actually it will sit in a
  loop waiting for UDP packets that will never arrive.)  This causes
  the NixOS reboot to hang, since Upstart can't kill jobs stuck in
  their start script.

svn path=/nixos/branches/modular-nixos/; revision=15829
2009-06-02 19:40:14 +00:00

109 lines
2.2 KiB
Nix

{pkgs, config, ...}:
###### interface
let
inherit (pkgs.lib) mkOption mkIf;
options = {
services = {
ntp = {
enable = mkOption {
default = true;
description = "
Whether to synchronise your machine's time using the NTP
protocol.
";
};
servers = mkOption {
default = [
"0.pool.ntp.org"
"1.pool.ntp.org"
"2.pool.ntp.org"
];
description = "
The set of NTP servers from which to synchronise.
";
};
};
};
};
in
###### implementation
let
inherit (pkgs) writeText ntp;
stateDir = "/var/lib/ntp";
ntpUser = "ntp";
servers = config.services.ntp.servers;
modprobe = config.system.sbin.modprobe;
configFile = writeText "ntp.conf" ''
driftfile ${stateDir}/ntp.drift
# Keep the drift file in ${stateDir}/ntp.drift. However, since we
# chroot to ${stateDir}, we have to specify it as /ntp.drift.
driftfile /ntp.drift
${toString (map (server: "server " + server + "\n") servers)}
'';
ntpFlags = "-c ${configFile} -u ${ntpUser}:nogroup -i ${stateDir}";
in
mkIf config.services.ntp.enable {
require = [
options
];
services = {
extraJobs = [{
name = "ntpd";
users = [
{ name = ntpUser;
uid = config.ids.uids.ntp;
description = "NTP daemon user";
home = stateDir;
}
];
job = ''
description "NTP daemon"
start on ip-up
stop on ip-down
stop on shutdown
start script
mkdir -m 0755 -p ${stateDir}
chown ${ntpUser} ${stateDir}
# Needed to run ntpd as an unprivileged user.
${modprobe}/sbin/modprobe capability || true
# !!! This can hang indefinitely if the network is down or
# the servers are unreachable. This is particularly bad
# because Upstart cannot kill jobs stuck in the start
# phase. Thus a hanging ntpd job can block system
# shutdown.
# ${ntp}/bin/ntpd -q -g ${ntpFlags}
end script
respawn ${ntp}/bin/ntpd -g -n ${ntpFlags}
'';
}];
};
}