Run-time reconfiguration

Knot Resolver offers several ways to modify its configuration at run-time:

  • Using control socket driven by an external system

  • Using Lua program embedded in Resolver’s configuration file

Both ways can also be combined: For example the configuration file can contain a little Lua function which gathers statistics and returns them in JSON string. This can be used by an external system which uses control socket to call this user-defined function and to retrieve its results.

Control sockets

Control socket acts like “an interactive configuration file” so all actions available in configuration file can be executed interactively using the control socket. One possible use-case is reconfiguring the resolver instances from another program, e.g. a maintenance script.


Each instance of Knot Resolver exposes its own control socket. Take that into account when scripting deployments with Multiple instances.

When Knot Resolver is started using Systemd (see section Startup) it creates a control socket in path /run/knot-resolver/control/$ID. Connection to the socket can be made from command line using e.g. socat:

$ socat - UNIX-CONNECT:/run/knot-resolver/control/1

When successfully connected to a socket, the command line should change to something like >. Then you can interact with kresd to see configuration or set a new one. There are some basic commands to start with.

> help()            -- shows help
> net.interfaces()  -- lists available interfaces
> net.list()        -- lists running network services

The direct output of commands sent over socket is captured and sent back, which gives you an immediate response on the outcome of your command. The commands and their output are also logged in contrl group, on debug level if successful or warning level if failed (see around log_level()).

Control sockets are also a way to enumerate and test running instances, the list of sockets corresponds to the list of processes, and you can test the process for liveliness by connecting to the UNIX socket.


Executes the provided string as lua code on every running resolver instance and returns the results as a table.

Key n is always present in the returned table and specifies the total number of instances the command was executed on. The table also contains results from each instance accessible through keys 1 to n (inclusive). If any instance returns nil, it is not explicitly part of the table, but you can detect it by iterating through 1 to n.

> map('')  -- return an ID of every active instance
    ['n'] = 2,
> map(' == "1" or nil')  -- example of `nil` return value
    [2] = true,
    ['n'] = 2,

The order of instances isn’t guaranteed or stable. When you need to identify the instances, you may use kluautil.kr_table_pack() function to return multiple values as a table. It uses similar semantics with n as described above to allow nil values.

> map('require("kluautil").kr_table_pack(, stats.get(""))')
        ['n'] = 2,
        ['n'] = 2,
    ['n'] = 2,

If the command fails on any instance, an error is returned and the execution is in an undefined state (the command might not have been executed on all instances). When using the map() function to execute any code that might fail, your code should be wrapped in pcall() to avoid this issue.

> map('require("kluautil").kr_table_pack(pcall(net.tls, "cert.pem", "key.pem"))')
        true,  -- function succeeded
        true,  -- function return value(s)
        ['n'] = 2,
        false,  -- function failed
        'error occurred...',  -- the returned error message
        ['n'] = 2,
    ['n'] = 2,

Lua scripts

As it was mentioned in section Syntax, Resolver’s configuration file contains program in Lua programming language. This allows you to write dynamic rules and helps you to avoid repetitive templating that is unavoidable with static configuration. For example parts of configuration can depend on hostname() of the machine:

if hostname() == 'hidden' then
        net.listen(net.eth0, 5353)

Another example would show how it is possible to bind to all interfaces, using iteration.

for name, addr_list in pairs(net.interfaces()) do


Some users observed a considerable, close to 100%, performance gain in Docker containers when they bound the daemon to a single interface:ip address pair. One may expand the aforementioned example with browsing available addresses as:

for k, v in pairs(addr_list["addr"]) do
        if string.sub(v,1,string.len(addrpref)) == addrpref then

You can also use third-party Lua libraries (available for example through LuaRocks) as on this example to download cache from parent, to avoid cold-cache start.

local http = require('socket.http')
local ltn12 = require('ltn12')

local cache_size = 100*MB
local cache_path = '/var/cache/knot-resolver', 'lmdb://' .. cache_path)
if cache.count() == 0 then
        -- download cache from parent
        http.request {
                url = 'http://parent/data.mdb',
                sink = ltn12.sink.file( .. '/data.mdb', 'w'))
        -- reopen cache with 100M limit, 'lmdb://' .. cache_path)

Helper functions

Following built-in functions are useful for scripting:

env (table)

Retrieve environment variables.


env.USER -- equivalent to $USER in shell

Lua representation of data in JSON string.


> fromjson('{"key1": "value1", "key2": {"subkey1": 1, "subkey2": 2}}')
[key1] => value1
[key2] => {
    [subkey1] => 1
    [subkey2] => 2

Machine hostname.

If called with a parameter, it will set kresd’s internal hostname. If called without a parameter, it will return kresd’s internal hostname, or the system’s POSIX hostname (see gethostname(2)) if kresd’s internal hostname is unset.

This also affects ephemeral (self-signed) certificates generated by kresd for DNS over TLS.


Current package version as string.


> package_version()
resolve(name, type[, class = kres.class.IN, options = {}, finish = nil, init = nil])
  • name (string) – Query name (e.g. ‘com.’)

  • type (number) – Query type (e.g. kres.type.NS)

  • class (number) – Query class (optional) (e.g. kres.class.IN)

  • options (strings) – Resolution options (see kr_qflags)

  • finish (function) – Callback to be executed when resolution completes (e.g. function cb (pkt, req) end). The callback gets a packet containing the final answer and doesn’t have to return anything.

  • init (function) – Callback to be executed with the kr_request before resolution starts.


boolean, true if resolution was started

The function can also be executed with a table of arguments instead. This is useful if you’d like to skip some arguments, for example:

resolve {
   name = '',
   type = kres.type.AAAA,
   init = function (req)


-- Send query for root DNSKEY, ignore cache
resolve('.', kres.type.DNSKEY, kres.class.IN, 'NO_CACHE')

-- Query for AAAA record
resolve('', kres.type.AAAA, kres.class.IN, 0,
function (pkt, req)
   -- Check answer RCODE
   if pkt:rcode() == kres.rcode.NOERROR then
      -- Print matching records
      local records = pkt:section(kres.section.ANSWER)
      for i = 1, #records do
         local rr = records[i]
         if rr.type == kres.type.AAAA then
            print ('record:', kres.rr2str(rr))
      print ('rcode: ', pkt:rcode())

JSON text representation of object.


> testtable = { key1 = "value1", "key2" = { subkey1 = 1, subkey2 = 2 } }
> tojson(testtable)

Asynchronous events

Lua language used in configuration file allows you to script actions upon various events, for example publish statistics each minute. Following example uses built-in function event.recurrent() which calls user-supplied anonymous function:

local ffi = require('ffi')

-- log statistics every second
local stat_id = event.recurrent(1 * second, function(evid)
     log_info(ffi.C.LOG_GRP_STATISTICS, table_print(stats.list()))

-- stop printing statistics after first minute
event.after(1 * minute, function(evid)

Note that each scheduled event is identified by a number valid for the duration of the event, you may use it to cancel the event at any time.

To persist state between two invocations of a function Lua uses concept called closures. In the following example function speed_monitor() is a closure function, which provides persistent variable called previous.

local ffi = require('ffi')

-- make a closure, encapsulating counter
function speed_monitor()
    local previous = stats.list()
    -- monitoring function
    return function(evid)
        local now = stats.list()
        local total_increment = now[''] - previous['']
        local slow_increment = now['answer.slow'] - previous['answer.slow']
        if slow_increment / total_increment > 0.05 then
            log_warn(ffi.C.LOG_GRP_STATISTICS, 'WARNING! More than 5 %% of queries was slow!')
        previous = now  -- store current value in closure

 -- monitor every minute
 local monitor_id = event.recurrent(1 * minute, speed_monitor())

Another type of actionable event is activity on a file descriptor. This allows you to embed other event loops or monitor open files and then fire a callback when an activity is detected. This allows you to build persistent services like monitoring probes that cooperate well with the daemon internal operations. See event.socket().

Filesystem watchers are possible with worker.coroutine() and cqueues, see the cqueues documentation for more information. Here is an simple example:

local notify = require('cqueues.notify')
local watcher = notify.opendir('/etc')

-- Watch changes to /etc/hosts
worker.coroutine(function ()
  for flags, name in watcher:changes() do
    for flag in notify.flags(flags) do
      -- print information about the modified file
      print(name, notify[flag])

Timers and events reference

The timer represents exactly the thing described in the examples - it allows you to execute closures after specified time, or event recurrent events. Time is always described in milliseconds, but there are convenient variables that you can use - sec, minute, hour. For example, 5 * hour represents five hours, or 5*60*60*100 milliseconds.

event.after(time, function)

event id

Execute function after the specified time has passed. The first parameter of the callback is the event itself.


event.after(1 * minute, function() print('Hi!') end)
event.recurrent(interval, function)

event id

Execute function immediately and then periodically after each interval.


msg_count = 0
event.recurrent(5 * sec, function(e)
   msg_count = msg_count + 1
   print('Hi #'..msg_count)
event.reschedule(event_id, timeout)

Reschedule a running event, it has no effect on canceled events. New events may reuse the event_id, so the behaviour is undefined if the function is called after another event is started.


local interval = 1 * minute
event.after(1 * minute, function (ev)
   print('Good morning!')
   -- Halve the interval for each iteration
   interval = interval / 2
   event.reschedule(ev, interval)

Cancel running event, it has no effect on already canceled events. New events may reuse the event_id, so the behaviour is undefined if the function is called after another event is started.


e = event.after(1 * minute, function() print('Hi!') end)

Watch for file descriptor activity. This allows embedding other event loops or simply firing events when a pipe endpoint becomes active. In another words, asynchronous notifications for daemon.

event.socket(fd, cb)
  • fd (number) – file descriptor to watch

  • cb – closure or callback to execute when fd becomes active


event id

Execute function when there is activity on the file descriptor and calls a closure with event id as the first parameter, status as second and number of events as third.


e = event.socket(0, function(e, status, nevents)
   print('activity detected')

Asynchronous function execution

The event package provides a very basic mean for non-blocking execution - it allows running code when activity on a file descriptor is detected, and when a certain amount of time passes. It doesn’t however provide an easy to use abstraction for non-blocking I/O. This is instead exposed through the worker package (if cqueues Lua package is installed in the system).


Start a new coroutine with given function (closure). The function can do I/O or run timers without blocking the main thread. See cqueues for documentation of possible operations and synchronization primitives. The main limitation is that you can’t wait for a finish of a coroutine from processing layers, because it’s not currently possible to suspend and resume execution of processing layers.


worker.coroutine(function ()
  for i = 0, 10 do
    print('executing', i)

Pause execution of current function (asynchronously if running inside a worker coroutine).


function async_print(testname, sleep)
        log(testname .. ': system time before sleep' .. tostring(os.time())
        worker.sleep(sleep)  -- other coroutines continue execution now
        log(testname .. ': system time AFTER sleep' .. tostring(os.time())

worker.coroutine(function() async_print('call #1', 5) end)
worker.coroutine(function() async_print('call #2', 3) end)

Output from this example demonstrates that both calls to function async_print were executed asynchronously:

call #2: system time before sleep 1578065073
call #1: system time before sleep 1578065073
call #2: system time AFTER  sleep 1578065076
call #1: system time AFTER  sleep 1578065078

Etcd support

The etcd module connects to etcd peers and watches for configuration changes. By default, the module watches the subtree under /knot-resolver directory, but you can change this in the etcd library configuration.

The subtree structure corresponds to the configuration variables in the declarative style.

$ etcdctl set /knot-resolver/net/ 53
$ etcdctl set /knot-resolver/cache/size 10000000

Configures all listening nodes to following configuration:

net = { '' }
cache.size = 10000000

Example configuration

        prefix = '/knot-resolver',
        peer = ''


Work in progress!


  • lua-etcd library available in LuaRocks

    $ luarocks --lua-version 5.1 install etcd --from=