OpenBSD manual page server

Manual Page Search Parameters
nsd-control(8) nsd 4.9.1 nsd-control(8)

nsd-control, nsd-control-setup - NSD remote server control utility.

nsd-control [-c cfgfile] [-s server] command

nsd-control performs remote administration on the nsd(8) DNS server. It reads the configuration file, contacts the nsd server over SSL, sends the command and displays the result. Commands that require a reload are queued and the result indicates the command was accepted.

The available options are:

Show the version and commandline option help.
The config file to read with settings. If not given the default config file /var/nsd/etc/nsd.conf is used.
IPv4 or IPv6 address of the server to contact. If not given, the address is read from the config file.

There are several commands that the server understands.

Start the server. Simply execs nsd(8). The nsd executable is not searched for in the PATH set in the environment. Instead the default location relative to the installation prefix specified at compile-time. The executable location can be overridden by setting NSD_PATH in the environment. It is started with the config file specified using -c or the default config file.
Stop the server. The server daemon exits.
Reload zonefiles and reopen logfile. Without argument reads changed zonefiles. With argument reads the zonefile for the given zone and loads it.
Reload nsd.conf and apply changes to TSIG keys and configuration patterns, and apply the changes to add and remove zones that are mentioned in the config. Other changes are not applied, such as listening ip address and port and chroot, also per-zone statistics are not applied. The pattern updates means that the configuration options for zones (request-xfr, zonefile, notify, ...) are updated. Also new patterns are available for use with the addzone command.
Same as the reconfig option.
Reopen the logfile, for log rotate that wants to move the logfile away and create a new logfile. The log can also be reopened with kill -HUP (which also reloads all zonefiles).
Display server status. Exit code 3 if not running (the connection to the port is refused), 1 on error, 0 if running.
Output a sequence of name=value lines with statistics information, requires NSD to be compiled with this option enabled.
Same as stats, but does not zero the counters.
Add a new zone to the running server. The zone is added to the zonelist file on disk, so it stays after a restart. The pattern name determines the options for the new zone. For secondary zones a zone transfer is immediately attempted. For zones with a zonefile, the zone file is attempted to be read in.
Remove the zone from the running server. The zone is removed from the zonelist file on disk, from the nsd.db file and from the memory. If it had a zonefile, this remains (but may be outdated). Zones configured inside nsd.conf itself cannot be removed this way because the daemon does not write to the nsd.conf file, you need to add such zones to the zonelist file to be able to delete them with the delzone command.
Change a zone to use the pattern for options. The zone is deleted and added in one operation, changing it to use the new pattern for the zone options. Zones configured in nsd.conf cannot be changed like this, instead edit the nsd.conf (or the included file in nsd.conf) and reconfig.
Add zones read from stdin of nsd-control. Input is read per line, with name space patternname on a line. For bulk additions.
Remove zones read from stdin of nsd-control. Input is one name per line. For bulk removals.
Write zonefiles to disk, or the given zonefile to disk. Zones that have changed (via AXFR or IXFR) are written, or if the zonefile has not been created yet then it is created. Directory components of the zonefile path are created if necessary. With argument that zone is written if it was modified, without argument, all modified zones are written.
Send NOTIFY messages to secondary servers. Sends to the IP addresses configured in the 'notify:' lists for the primary zones hosted on this server. Usually NSD sends NOTIFY messages right away when a primary zone serial is updated. If a zone is given, notifies are sent for that zone. These secondary servers are supposed to initiate a zone transfer request later (to this server or another primary), this can be allowed via the 'provide-xfr:' acl list configuration. With argument that zone is processed, without argument, all zones are processed.
Attempt to update secondary zones that are hosted on this server by contacting the primaries. The primaries are configured via 'request-xfr:' lists. If a zone is given, that zone is updated. Usually NSD receives a NOTIFY from the primaries (configured via 'allow-notify:' acl list) that a new zone serial has to be transferred. For zones with no content, NSD may have backed off from asking often because the primaries did not respond, but this command will reset the backoff to its initial timeout, for frequent retries. With argument that zone is transferred, without argument, all zones are transferred.
Force update secondary zones that are hosted on this server. Even if the primary hosts the same serial number of the zone, a full AXFR is performed to fetch it. If you want to use IXFR and check that the serial number increases, use the 'transfer' command. With argument that zone is transferred, without argument, all zones are transferred.
Print state of the zone, the serial numbers and since when they have been acquired. Also prints the notify action (to which server), and zone transfer (and from which primary) if there is activity right now. The state of the zone is printed as: 'primary' (primary zones), 'ok' (secondary zone is up-to-date), 'expired' (secondary zone has expired), 'refreshing' (secondary zone has transfers active). The serial numbers printed are the 'served-serial' (currently active), the 'commit-serial' (is in reload), the 'notified-serial' (got notify, busy fetching the data). The serial numbers are only printed if such a serial number is available. With argument that zone is printed, without argument, all zones are printed.
Prints the PID of the server process. This is used for statistics (and only works when NSD is compiled with statistics enabled). This pid is not for sending unix signals, use the pid from nsd.pid for that, that pid is also stable.
Change logging verbosity.
print the secret and algorithm for the TSIG key with that name. Or list all the tsig keys with their name, secret and algorithm.
Change existing TSIG key with name to the new secret. The secret is a base64 encoded string. The changes are only in-memory and are gone next restart, for lasting changes edit the nsd.conf file or a file included from it.
Add a new TSIG key with the given name, secret and algorithm. Without algorithm a default (hmac-sha256) algorithm is used. The secret is a base64 encoded string. The changes are only in-memory and are gone next restart, for lasting changes edit the nsd.conf file or a file included from it.
Associate the zone with the given tsig. The access control lists for notify, allow-notify, provide-xfr and request-xfr are adjusted to use the given key.
Delete the TSIG key with the given name. Prints error if the key is still in use by some zone. The changes are only in-memory and are gone next restart, for lasting changes edit the nsd.conf file or a file included from it.
Add or replace a cookie secret persistently. <secret> needs to be an 128 bit hex string.

Cookie secrets can be either active or staging. Active cookie secrets are used to create DNS Cookies, but verification of a DNS Cookie succeeds with any of the active or staging cookie secrets. The state of the current cookie secrets can be printed with the print_cookie_secrets command.

When there are no cookie secrets configured yet, the <secret> is added as active. If there is already an active cookie secret, the <secret> is added as staging or replacing an existing staging secret.

To "roll" a cookie secret used in an anycast set. The new secret has to be added as staging secret to all nodes in the anycast set. When all nodes can verify DNS Cookies with the new secret, the new secret can be activated with the activate_cookie_secret command. After all nodes have the new secret active for at least one hour, the previous secret can be dropped with the drop_cookie_secret command.

Persistence is accomplished by writing to a file which if configured with the cookie-secret-file option in the server section of the config file. The default value for that is: /var/nsd/etc/nsd_cookiesecrets.txt .

Drop the staging cookie secret.
Make the current staging cookie secret active, and the current active cookie secret staging.
Show the current configured cookie secrets with their status.

The nsd-control program exits with status code 1 on error, 0 on success.

The setup requires a self-signed certificate and private keys for both the server and client. The script nsd-control-setup generates these in the default run directory, or with -d in another directory. If you change the access control permissions on the key files you can decide who can use nsd-control, by default owner and group but not all users. The script preserves private keys present in the directory. After running the script as root, turn on control-enable in nsd.conf.

The stats command shows a number of statistic counters.

number of queries received (the tls, tcp and udp queries added up).
number of queries handled by the server process. The number of server processes is set with the config statement server-count.
uptime in seconds since the server was started. With fractional seconds.
time since the last stats report, in seconds. With fractional seconds. Can be zero if polled quickly and the previous stats command resets the counters, so that the next gets a fully zero, and zero elapsed time, report.
size of nsd.db on disk, in bytes.
size of the DNS database in memory, in bytes.
size of memory for zone transfers and notifies in xfrd process, excludes TSIG data, in bytes.
size of zonelist file on disk, excludes the nsd.conf size, in bytes.
size of config data in memory, kept twice in server and xfrd process, in bytes.
number of queries with this query type.
number of queries with this opcode.
number of queries with this query class.
number of answers that carried this return code.
number of queries with EDNS OPT.
number of queries which failed EDNS parse.
number of queries over UDP ip4.
number of queries over UDP ip6.
number of connections over TCP ip4.
number of connections over TCP ip6.
number of connections over TLS ip4. TLS queries are not part of num.tcp.
number of connections over TLS ip6. TLS queries are not part of num.tcp6.
number of answers with NOERROR rcode and without AA flag, this includes the referrals.
number of queries for which the receive failed.
number of answers for which the transmit failed.
number of AXFR requests from clients (that got served with reply).
number of IXFR requests from clients (that got served with reply).
number of answers with TC flag set.
number of queries that were dropped because they failed sanity check.
number of primary zones served. These are zones with no 'request-xfr:' entries. Also output as 'zone.master' for backwards compatibility.
number of secondary zones served. These are zones with 'request-xfr' entries. Also output as 'zone.slave' for backwards compatibility.

/var/nsd/etc/nsd.conf
nsd configuration file.
/var/nsd/etc
directory with private keys (nsd_server.key and nsd_control.key) and self-signed certificates (nsd_server.pem and nsd_control.pem).

nsd.conf(5), nsd(8), nsd-checkconf(8)

April 4, 2024 NLnet Labs