nixpkgs/nixos/modules/services/networking/nsd.nix
Félix Baylac-Jacqué 7020dc8eac
nixos/nsd: symlink conf file to /etc/nsd
We remove the configFile build flag override in the NixOS module.

Instead of embedding the conf file link to the binaries, we symlink it
to /etc/nsd/nsd.nix, the hardcoded config file location for the
various CLI nsd utilities.

This config file build option override is triggerring a nsd rebuild
for each configuration change. This prevent us to use the nixos cache
in many cases.

Co-authored-by: Erjo <erjo@cocoba.work>
2020-06-26 20:18:33 +02:00

982 lines
29 KiB
Nix

{ config, pkgs, lib, ... }:
with lib;
let
cfg = config.services.nsd;
username = "nsd";
stateDir = "/var/lib/nsd";
pidFile = stateDir + "/var/nsd.pid";
# build nsd with the options needed for the given config
nsdPkg = pkgs.nsd.override {
bind8Stats = cfg.bind8Stats;
ipv6 = cfg.ipv6;
ratelimit = cfg.ratelimit.enable;
rootServer = cfg.rootServer;
zoneStats = length (collect (x: (x.zoneStats or null) != null) cfg.zones) > 0;
};
mkZoneFileName = name: if name == "." then "root" else name;
nsdEnv = pkgs.buildEnv {
name = "nsd-env";
paths = [ configFile ]
++ mapAttrsToList (name: zone: writeZoneData name zone.data) zoneConfigs;
postBuild = ''
echo "checking zone files"
cd $out/zones
for zoneFile in *; do
echo "|- checking zone '$out/zones/$zoneFile'"
${nsdPkg}/sbin/nsd-checkzone "$zoneFile" "$zoneFile" || {
if grep -q \\\\\\$ "$zoneFile"; then
echo zone "$zoneFile" contains escaped dollar signes \\\$
echo Escaping them is not needed any more. Please make shure \
to unescape them where they prefix a variable name
fi
exit 1
}
done
echo "checking configuration file"
${nsdPkg}/sbin/nsd-checkconf $out/nsd.conf
'';
};
writeZoneData = name: text: pkgs.writeTextFile {
name = "nsd-zone-${mkZoneFileName name}";
inherit text;
destination = "/zones/${mkZoneFileName name}";
};
# options are ordered alphanumerically by the nixos option name
configFile = pkgs.writeTextDir "nsd.conf" ''
server:
chroot: "${stateDir}"
username: ${username}
# The directory for zonefile: files. The daemon chdirs here.
zonesdir: "${stateDir}"
# the list of dynamically added zones.
database: "${stateDir}/var/nsd.db"
pidfile: "${pidFile}"
xfrdfile: "${stateDir}/var/xfrd.state"
xfrdir: "${stateDir}/tmp"
zonelistfile: "${stateDir}/var/zone.list"
# interfaces
${forEach " ip-address: " cfg.interfaces}
ip-freebind: ${yesOrNo cfg.ipFreebind}
hide-version: ${yesOrNo cfg.hideVersion}
identity: "${cfg.identity}"
ip-transparent: ${yesOrNo cfg.ipTransparent}
do-ip4: ${yesOrNo cfg.ipv4}
ipv4-edns-size: ${toString cfg.ipv4EDNSSize}
do-ip6: ${yesOrNo cfg.ipv6}
ipv6-edns-size: ${toString cfg.ipv6EDNSSize}
log-time-ascii: ${yesOrNo cfg.logTimeAscii}
${maybeString "nsid: " cfg.nsid}
port: ${toString cfg.port}
reuseport: ${yesOrNo cfg.reuseport}
round-robin: ${yesOrNo cfg.roundRobin}
server-count: ${toString cfg.serverCount}
${maybeToString "statistics: " cfg.statistics}
tcp-count: ${toString cfg.tcpCount}
tcp-query-count: ${toString cfg.tcpQueryCount}
tcp-timeout: ${toString cfg.tcpTimeout}
verbosity: ${toString cfg.verbosity}
${maybeString "version: " cfg.version}
xfrd-reload-timeout: ${toString cfg.xfrdReloadTimeout}
zonefiles-check: ${yesOrNo cfg.zonefilesCheck}
${maybeString "rrl-ipv4-prefix-length: " cfg.ratelimit.ipv4PrefixLength}
${maybeString "rrl-ipv6-prefix-length: " cfg.ratelimit.ipv6PrefixLength}
rrl-ratelimit: ${toString cfg.ratelimit.ratelimit}
${maybeString "rrl-slip: " cfg.ratelimit.slip}
rrl-size: ${toString cfg.ratelimit.size}
rrl-whitelist-ratelimit: ${toString cfg.ratelimit.whitelistRatelimit}
${keyConfigFile}
remote-control:
control-enable: ${yesOrNo cfg.remoteControl.enable}
control-key-file: "${cfg.remoteControl.controlKeyFile}"
control-cert-file: "${cfg.remoteControl.controlCertFile}"
${forEach " control-interface: " cfg.remoteControl.interfaces}
control-port: ${toString cfg.remoteControl.port}
server-key-file: "${cfg.remoteControl.serverKeyFile}"
server-cert-file: "${cfg.remoteControl.serverCertFile}"
${concatStrings (mapAttrsToList zoneConfigFile zoneConfigs)}
${cfg.extraConfig}
'';
yesOrNo = b: if b then "yes" else "no";
maybeString = prefix: x: if x == null then "" else ''${prefix} "${x}"'';
maybeToString = prefix: x: if x == null then "" else ''${prefix} ${toString x}'';
forEach = pre: l: concatMapStrings (x: pre + x + "\n") l;
keyConfigFile = concatStrings (mapAttrsToList (keyName: keyOptions: ''
key:
name: "${keyName}"
algorithm: "${keyOptions.algorithm}"
include: "${stateDir}/private/${keyName}"
'') cfg.keys);
copyKeys = concatStrings (mapAttrsToList (keyName: keyOptions: ''
secret=$(cat "${keyOptions.keyFile}")
dest="${stateDir}/private/${keyName}"
echo " secret: \"$secret\"" > "$dest"
chown ${username}:${username} "$dest"
chmod 0400 "$dest"
'') cfg.keys);
# options are ordered alphanumerically by the nixos option name
zoneConfigFile = name: zone: ''
zone:
name: "${name}"
zonefile: "${stateDir}/zones/${mkZoneFileName name}"
${maybeString "outgoing-interface: " zone.outgoingInterface}
${forEach " rrl-whitelist: " zone.rrlWhitelist}
${maybeString "zonestats: " zone.zoneStats}
${maybeToString "max-refresh-time: " zone.maxRefreshSecs}
${maybeToString "min-refresh-time: " zone.minRefreshSecs}
${maybeToString "max-retry-time: " zone.maxRetrySecs}
${maybeToString "min-retry-time: " zone.minRetrySecs}
allow-axfr-fallback: ${yesOrNo zone.allowAXFRFallback}
${forEach " allow-notify: " zone.allowNotify}
${forEach " request-xfr: " zone.requestXFR}
${forEach " notify: " zone.notify}
notify-retry: ${toString zone.notifyRetry}
${forEach " provide-xfr: " zone.provideXFR}
'';
zoneConfigs = zoneConfigs' {} "" { children = cfg.zones; };
zoneConfigs' = parent: name: zone:
if !(zone ? children) || zone.children == null || zone.children == { }
# leaf -> actual zone
then listToAttrs [ (nameValuePair name (parent // zone)) ]
# fork -> pattern
else zipAttrsWith (name: head) (
mapAttrsToList (name: child: zoneConfigs' (parent // zone // { children = {}; }) name child)
zone.children
);
# fighting infinite recursion
zoneOptions = zoneOptionsRaw // childConfig zoneOptions1 true;
zoneOptions1 = zoneOptionsRaw // childConfig zoneOptions2 false;
zoneOptions2 = zoneOptionsRaw // childConfig zoneOptions3 false;
zoneOptions3 = zoneOptionsRaw // childConfig zoneOptions4 false;
zoneOptions4 = zoneOptionsRaw // childConfig zoneOptions5 false;
zoneOptions5 = zoneOptionsRaw // childConfig zoneOptions6 false;
zoneOptions6 = zoneOptionsRaw // childConfig null false;
childConfig = x: v: { options.children = { type = types.attrsOf x; visible = v; }; };
# options are ordered alphanumerically
zoneOptionsRaw = types.submodule {
options = {
allowAXFRFallback = mkOption {
type = types.bool;
default = true;
description = ''
If NSD as secondary server should be allowed to AXFR if the primary
server does not allow IXFR.
'';
};
allowNotify = mkOption {
type = types.listOf types.str;
default = [ ];
example = [ "192.0.2.0/24 NOKEY" "10.0.0.1-10.0.0.5 my_tsig_key_name"
"10.0.3.4&255.255.0.0 BLOCKED"
];
description = ''
Listed primary servers are allowed to notify this secondary server.
<screen><![CDATA[
Format: <ip> <key-name | NOKEY | BLOCKED>
<ip> either a plain IPv4/IPv6 address or range. Valid patters for ranges:
* 10.0.0.0/24 # via subnet size
* 10.0.0.0&255.255.255.0 # via subnet mask
* 10.0.0.1-10.0.0.254 # via range
A optional port number could be added with a '@':
* 2001:1234::1@1234
<key-name | NOKEY | BLOCKED>
* <key-name> will use the specified TSIG key
* NOKEY no TSIG signature is required
* BLOCKED notifies from non-listed or blocked IPs will be ignored
* ]]></screen>
'';
};
children = mkOption {
default = {};
description = ''
Children zones inherit all options of their parents. Attributes
defined in a child will overwrite the ones of its parent. Only
leaf zones will be actually served. This way it's possible to
define maybe zones which share most attributes without
duplicating everything. This mechanism replaces nsd's patterns
in a save and functional way.
'';
};
data = mkOption {
type = types.lines;
default = "";
example = "";
description = ''
The actual zone data. This is the content of your zone file.
Use imports or pkgs.lib.readFile if you don't want this data in your config file.
'';
};
dnssec = mkEnableOption "DNSSEC";
dnssecPolicy = {
algorithm = mkOption {
type = types.str;
default = "RSASHA256";
description = "Which algorithm to use for DNSSEC";
};
keyttl = mkOption {
type = types.str;
default = "1h";
description = "TTL for dnssec records";
};
coverage = mkOption {
type = types.str;
default = "1y";
description = ''
The length of time to ensure that keys will be correct; no action will be taken to create new keys to be activated after this time.
'';
};
zsk = mkOption {
type = keyPolicy;
default = { keySize = 2048;
prePublish = "1w";
postPublish = "1w";
rollPeriod = "1mo";
};
description = "Key policy for zone signing keys";
};
ksk = mkOption {
type = keyPolicy;
default = { keySize = 4096;
prePublish = "1mo";
postPublish = "1mo";
rollPeriod = "0";
};
description = "Key policy for key signing keys";
};
};
maxRefreshSecs = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
Limit refresh time for secondary zones. This is the timer which
checks to see if the zone has to be refetched when it expires.
Normally the value from the SOA record is used, but this option
restricts that value.
'';
};
minRefreshSecs = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
Limit refresh time for secondary zones.
'';
};
maxRetrySecs = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
Limit retry time for secondary zones. This is the timeout after
a failed fetch attempt for the zone. Normally the value from
the SOA record is used, but this option restricts that value.
'';
};
minRetrySecs = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
Limit retry time for secondary zones.
'';
};
notify = mkOption {
type = types.listOf types.str;
default = [];
example = [ "10.0.0.1@3721 my_key" "::5 NOKEY" ];
description = ''
This primary server will notify all given secondary servers about
zone changes.
<screen><![CDATA[
Format: <ip> <key-name | NOKEY>
<ip> a plain IPv4/IPv6 address with on optional port number (ip@port)
<key-name | NOKEY>
* <key-name> sign notifies with the specified key
* NOKEY don't sign notifies
]]></screen>
'';
};
notifyRetry = mkOption {
type = types.int;
default = 5;
description = ''
Specifies the number of retries for failed notifies. Set this along with notify.
'';
};
outgoingInterface = mkOption {
type = types.nullOr types.str;
default = null;
example = "2000::1@1234";
description = ''
This address will be used for zone-transfere requests if configured
as a secondary server or notifications in case of a primary server.
Supply either a plain IPv4 or IPv6 address with an optional port
number (ip@port).
'';
};
provideXFR = mkOption {
type = types.listOf types.str;
default = [];
example = [ "192.0.2.0/24 NOKEY" "192.0.2.0/24 my_tsig_key_name" ];
description = ''
Allow these IPs and TSIG to transfer zones, addr TSIG|NOKEY|BLOCKED
address range 192.0.2.0/24, 1.2.3.4&amp;255.255.0.0, 3.0.2.20-3.0.2.40
'';
};
requestXFR = mkOption {
type = types.listOf types.str;
default = [];
example = [];
description = ''
Format: <code>[AXFR|UDP] &lt;ip-address&gt; &lt;key-name | NOKEY&gt;</code>
'';
};
rrlWhitelist = mkOption {
type = with types; listOf (enum [ "nxdomain" "error" "referral" "any" "rrsig" "wildcard" "nodata" "dnskey" "positive" "all" ]);
default = [];
description = ''
Whitelists the given rrl-types.
'';
};
zoneStats = mkOption {
type = types.nullOr types.str;
default = null;
example = "%s";
description = ''
When set to something distinct to null NSD is able to collect
statistics per zone. All statistics of this zone(s) will be added
to the group specified by this given name. Use "%s" to use the zones
name as the group. The groups are output from nsd-control stats
and stats_noreset.
'';
};
};
};
keyPolicy = types.submodule {
options = {
keySize = mkOption {
type = types.int;
description = "Key size in bits";
};
prePublish = mkOption {
type = types.str;
description = "How long in advance to publish new keys";
};
postPublish = mkOption {
type = types.str;
description = "How long after deactivation to keep a key in the zone";
};
rollPeriod = mkOption {
type = types.str;
description = "How frequently to change keys";
};
};
};
dnssecZones = (filterAttrs (n: v: if v ? dnssec then v.dnssec else false) zoneConfigs);
dnssec = dnssecZones != {};
dnssecTools = pkgs.bind.override { enablePython = true; };
signZones = optionalString dnssec ''
mkdir -p ${stateDir}/dnssec
chown ${username}:${username} ${stateDir}/dnssec
chmod 0600 ${stateDir}/dnssec
${concatStrings (mapAttrsToList signZone dnssecZones)}
'';
signZone = name: zone: ''
${dnssecTools}/bin/dnssec-keymgr -g ${dnssecTools}/bin/dnssec-keygen -s ${dnssecTools}/bin/dnssec-settime -K ${stateDir}/dnssec -c ${policyFile name zone.dnssecPolicy} ${name}
${dnssecTools}/bin/dnssec-signzone -S -K ${stateDir}/dnssec -o ${name} -O full -N date ${stateDir}/zones/${name}
${nsdPkg}/sbin/nsd-checkzone ${name} ${stateDir}/zones/${name}.signed && mv -v ${stateDir}/zones/${name}.signed ${stateDir}/zones/${name}
'';
policyFile = name: policy: pkgs.writeText "${name}.policy" ''
zone ${name} {
algorithm ${policy.algorithm};
key-size zsk ${toString policy.zsk.keySize};
key-size ksk ${toString policy.ksk.keySize};
keyttl ${policy.keyttl};
pre-publish zsk ${policy.zsk.prePublish};
pre-publish ksk ${policy.ksk.prePublish};
post-publish zsk ${policy.zsk.postPublish};
post-publish ksk ${policy.ksk.postPublish};
roll-period zsk ${policy.zsk.rollPeriod};
roll-period ksk ${policy.ksk.rollPeriod};
coverage ${policy.coverage};
};
'';
in
{
# options are ordered alphanumerically
options.services.nsd = {
enable = mkEnableOption "NSD authoritative DNS server";
bind8Stats = mkEnableOption "BIND8 like statistics";
dnssecInterval = mkOption {
type = types.str;
default = "1h";
description = ''
How often to check whether dnssec key rollover is required
'';
};
extraConfig = mkOption {
type = types.lines;
default = "";
description = ''
Extra nsd config.
'';
};
hideVersion = mkOption {
type = types.bool;
default = true;
description = ''
Whether NSD should answer VERSION.BIND and VERSION.SERVER CHAOS class queries.
'';
};
identity = mkOption {
type = types.str;
default = "unidentified server";
description = ''
Identify the server (CH TXT ID.SERVER entry).
'';
};
interfaces = mkOption {
type = types.listOf types.str;
default = [ "127.0.0.0" "::1" ];
description = ''
What addresses the server should listen to.
'';
};
ipFreebind = mkOption {
type = types.bool;
default = false;
description = ''
Whether to bind to nonlocal addresses and interfaces that are down.
Similar to ip-transparent.
'';
};
ipTransparent = mkOption {
type = types.bool;
default = false;
description = ''
Allow binding to non local addresses.
'';
};
ipv4 = mkOption {
type = types.bool;
default = true;
description = ''
Whether to listen on IPv4 connections.
'';
};
ipv4EDNSSize = mkOption {
type = types.int;
default = 4096;
description = ''
Preferred EDNS buffer size for IPv4.
'';
};
ipv6 = mkOption {
type = types.bool;
default = true;
description = ''
Whether to listen on IPv6 connections.
'';
};
ipv6EDNSSize = mkOption {
type = types.int;
default = 4096;
description = ''
Preferred EDNS buffer size for IPv6.
'';
};
logTimeAscii = mkOption {
type = types.bool;
default = true;
description = ''
Log time in ascii, if false then in unix epoch seconds.
'';
};
nsid = mkOption {
type = types.nullOr types.str;
default = null;
description = ''
NSID identity (hex string, or "ascii_somestring").
'';
};
port = mkOption {
type = types.int;
default = 53;
description = ''
Port the service should bind do.
'';
};
reuseport = mkOption {
type = types.bool;
default = pkgs.stdenv.isLinux;
description = ''
Whether to enable SO_REUSEPORT on all used sockets. This lets multiple
processes bind to the same port. This speeds up operation especially
if the server count is greater than one and makes fast restarts less
prone to fail
'';
};
rootServer = mkOption {
type = types.bool;
default = false;
description = ''
Whether this server will be a root server (a DNS root server, you
usually don't want that).
'';
};
roundRobin = mkEnableOption "round robin rotation of records";
serverCount = mkOption {
type = types.int;
default = 1;
description = ''
Number of NSD servers to fork. Put the number of CPUs to use here.
'';
};
statistics = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
Statistics are produced every number of seconds. Prints to log.
If null no statistics are logged.
'';
};
tcpCount = mkOption {
type = types.int;
default = 100;
description = ''
Maximum number of concurrent TCP connections per server.
'';
};
tcpQueryCount = mkOption {
type = types.int;
default = 0;
description = ''
Maximum number of queries served on a single TCP connection.
0 means no maximum.
'';
};
tcpTimeout = mkOption {
type = types.int;
default = 120;
description = ''
TCP timeout in seconds.
'';
};
verbosity = mkOption {
type = types.int;
default = 0;
description = ''
Verbosity level.
'';
};
version = mkOption {
type = types.nullOr types.str;
default = null;
description = ''
The version string replied for CH TXT version.server and version.bind
queries. Will use the compiled package version on null.
See hideVersion for enabling/disabling this responses.
'';
};
xfrdReloadTimeout = mkOption {
type = types.int;
default = 1;
description = ''
Number of seconds between reloads triggered by xfrd.
'';
};
zonefilesCheck = mkOption {
type = types.bool;
default = true;
description = ''
Whether to check mtime of all zone files on start and sighup.
'';
};
keys = mkOption {
type = types.attrsOf (types.submodule {
options = {
algorithm = mkOption {
type = types.str;
default = "hmac-sha256";
description = ''
Authentication algorithm for this key.
'';
};
keyFile = mkOption {
type = types.path;
description = ''
Path to the file which contains the actual base64 encoded
key. The key will be copied into "${stateDir}/private" before
NSD starts. The copied file is only accessibly by the NSD
user.
'';
};
};
});
default = {};
example = literalExample ''
{ "tsig.example.org" = {
algorithm = "hmac-md5";
keyFile = "/path/to/my/key";
};
}
'';
description = ''
Define your TSIG keys here.
'';
};
ratelimit = {
enable = mkEnableOption "ratelimit capabilities";
ipv4PrefixLength = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
IPv4 prefix length. Addresses are grouped by netblock.
'';
};
ipv6PrefixLength = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
IPv6 prefix length. Addresses are grouped by netblock.
'';
};
ratelimit = mkOption {
type = types.int;
default = 200;
description = ''
Max qps allowed from any query source.
0 means unlimited. With an verbosity of 2 blocked and
unblocked subnets will be logged.
'';
};
slip = mkOption {
type = types.nullOr types.int;
default = null;
description = ''
Number of packets that get discarded before replying a SLIP response.
0 disables SLIP responses. 1 will make every response a SLIP response.
'';
};
size = mkOption {
type = types.int;
default = 1000000;
description = ''
Size of the hashtable. More buckets use more memory but lower
the chance of hash hash collisions.
'';
};
whitelistRatelimit = mkOption {
type = types.int;
default = 2000;
description = ''
Max qps allowed from whitelisted sources.
0 means unlimited. Set the rrl-whitelist option for specific
queries to apply this limit instead of the default to them.
'';
};
};
remoteControl = {
enable = mkEnableOption "remote control via nsd-control";
controlCertFile = mkOption {
type = types.path;
default = "/etc/nsd/nsd_control.pem";
description = ''
Path to the client certificate signed with the server certificate.
This file is used by nsd-control and generated by nsd-control-setup.
'';
};
controlKeyFile = mkOption {
type = types.path;
default = "/etc/nsd/nsd_control.key";
description = ''
Path to the client private key, which is used by nsd-control
but not by the server. This file is generated by nsd-control-setup.
'';
};
interfaces = mkOption {
type = types.listOf types.str;
default = [ "127.0.0.1" "::1" ];
description = ''
Which interfaces NSD should bind to for remote control.
'';
};
port = mkOption {
type = types.int;
default = 8952;
description = ''
Port number for remote control operations (uses TLS over TCP).
'';
};
serverCertFile = mkOption {
type = types.path;
default = "/etc/nsd/nsd_server.pem";
description = ''
Path to the server self signed certificate, which is used by the server
but and by nsd-control. This file is generated by nsd-control-setup.
'';
};
serverKeyFile = mkOption {
type = types.path;
default = "/etc/nsd/nsd_server.key";
description = ''
Path to the server private key, which is used by the server
but not by nsd-control. This file is generated by nsd-control-setup.
'';
};
};
zones = mkOption {
type = types.attrsOf zoneOptions;
default = {};
example = literalExample ''
{ "serverGroup1" = {
provideXFR = [ "10.1.2.3 NOKEY" ];
children = {
"example.com." = {
data = '''
$ORIGIN example.com.
$TTL 86400
@ IN SOA a.ns.example.com. admin.example.com. (
...
''';
};
"example.org." = {
data = '''
$ORIGIN example.org.
$TTL 86400
@ IN SOA a.ns.example.com. admin.example.com. (
...
''';
};
};
};
"example.net." = {
provideXFR = [ "10.3.2.1 NOKEY" ];
data = '''
...
''';
};
}
'';
description = ''
Define your zones here. Zones can cascade other zones and therefore
inherit settings from parent zones. Look at the definition of
children to learn about inheritance and child zones.
The given example will define 3 zones (example.(com|org|net).). Both
example.com. and example.org. inherit their configuration from
serverGroup1.
'';
};
};
config = mkIf cfg.enable {
assertions = singleton {
assertion = zoneConfigs ? "." -> cfg.rootServer;
message = "You have a root zone configured. If this is really what you "
+ "want, please enable 'services.nsd.rootServer'.";
};
environment = {
systemPackages = [ nsdPkg ];
etc."nsd/nsd.conf".source = "${configFile}/nsd.conf";
};
users.groups.${username}.gid = config.ids.gids.nsd;
users.users.${username} = {
description = "NSD service user";
home = stateDir;
createHome = true;
uid = config.ids.uids.nsd;
group = username;
};
systemd.services.nsd = {
description = "NSD authoritative only domain name service";
after = [ "network.target" ];
wantedBy = [ "multi-user.target" ];
serviceConfig = {
ExecStart = "${nsdPkg}/sbin/nsd -d -c ${nsdEnv}/nsd.conf";
StandardError = "null";
PIDFile = pidFile;
Restart = "always";
RestartSec = "4s";
StartLimitBurst = 4;
StartLimitInterval = "5min";
};
preStart = ''
rm -Rf "${stateDir}/private/"
rm -Rf "${stateDir}/tmp/"
mkdir -m 0700 -p "${stateDir}/private"
mkdir -m 0700 -p "${stateDir}/tmp"
mkdir -m 0700 -p "${stateDir}/var"
cat > "${stateDir}/don't touch anything in here" << EOF
Everything in this directory except NSD's state in var and dnssec
is automatically generated and will be purged and redeployed by
the nsd.service pre-start script.
EOF
chown ${username}:${username} -R "${stateDir}/private"
chown ${username}:${username} -R "${stateDir}/tmp"
chown ${username}:${username} -R "${stateDir}/var"
rm -rf "${stateDir}/zones"
cp -rL "${nsdEnv}/zones" "${stateDir}/zones"
${copyKeys}
'';
};
systemd.timers.nsd-dnssec = mkIf dnssec {
description = "Automatic DNSSEC key rollover";
wantedBy = [ "nsd.service" ];
timerConfig = {
OnActiveSec = cfg.dnssecInterval;
OnUnitActiveSec = cfg.dnssecInterval;
};
};
systemd.services.nsd-dnssec = mkIf dnssec {
description = "DNSSEC key rollover";
wantedBy = [ "nsd.service" ];
before = [ "nsd.service" ];
script = signZones;
postStop = ''
/run/current-system/systemd/bin/systemctl kill -s SIGHUP nsd.service
'';
};
};
meta.maintainers = with lib.maintainers; [ hrdinka ];
}