» Forwarding DNS

By default, DNS is served from port 53. On most operating systems, this requires elevated privileges. Instead of running Consul with an administrative or root account, it is possible to instead forward appropriate queries to Consul, running on an unprivileged port, from another DNS server or port redirect.

In this guide, we will demonstrate forwarding from BIND as well as dnsmasq, Unbound, systemd-resolved, and iptables. For the sake of simplicity, BIND and Consul are running on the same machine in this example. For iptables the rules must be set on the same host as the Consul instance and relay hosts should not be on the same host or the redirects will intercept the traffic.

It is worth mentioning that, by default, Consul does not resolve DNS records outside the .consul. zone unless the recursors configuration option has been set. As an example of how this changes Consul's behavior, suppose a Consul DNS reply includes a CNAME record pointing outside the .consul TLD. The DNS reply will only include CNAME records by default. By contrast, when recursors is set and the upstream resolver is functioning correctly, Consul will try to resolve CNAMEs and include any records (e.g. A, AAAA, PTR) for them in its DNS reply.

You can either do one of the following:

» BIND Setup

First, you have to disable DNSSEC so that Consul and BIND can communicate. Here is an example of such a configuration:

options {
  listen-on port 53 { 127.0.0.1; };
  listen-on-v6 port 53 { ::1; };
  directory       "/var/named";
  dump-file       "/var/named/data/cache_dump.db";
  statistics-file "/var/named/data/named_stats.txt";
  memstatistics-file "/var/named/data/named_mem_stats.txt";
  allow-query     { localhost; };
  recursion yes;

  dnssec-enable no;
  dnssec-validation no;

  /* Path to ISC DLV key */
  bindkeys-file "/etc/named.iscdlv.key";

  managed-keys-directory "/var/named/dynamic";
};

include "/etc/named/consul.conf";

» Zone File

Then we set up a zone for our Consul managed records in consul.conf:

zone "consul" IN {
  type forward;
  forward only;
  forwarders { 127.0.0.1 port 8600; };
};

Here we assume Consul is running with default settings and is serving DNS on port 8600.

» Dnsmasq Setup

Dnsmasq is typically configured via a dnsmasq.conf or a series of files in the /etc/dnsmasq.d directory. In Dnsmasq's configuration file (e.g. /etc/dnsmasq.d/10-consul), add the following:

# Enable forward lookup of the 'consul' domain:
server=/consul/127.0.0.1#8600

# Uncomment and modify as appropriate to enable reverse DNS lookups for
# common netblocks found in RFC 1918, 5735, and 6598:
#rev-server=0.0.0.0/8,127.0.0.1#8600
#rev-server=10.0.0.0/8,127.0.0.1#8600
#rev-server=100.64.0.0/10,127.0.0.1#8600
#rev-server=127.0.0.1/8,127.0.0.1#8600
#rev-server=169.254.0.0/16,127.0.0.1#8600
#rev-server=172.16.0.0/12,127.0.0.1#8600
#rev-server=192.168.0.0/16,127.0.0.1#8600
#rev-server=224.0.0.0/4,127.0.0.1#8600
#rev-server=240.0.0.0/4,127.0.0.1#8600

Once that configuration is created, restart the dnsmasq service.

Additional useful settings in dnsmasq to consider include (see dnsmasq(8) for additional details):

# Accept DNS queries only from hosts whose address is on a local subnet.
#local-service

# Don't poll /etc/resolv.conf for changes.
#no-poll

# Don't read /etc/resolv.conf. Get upstream servers only from the command
# line or the dnsmasq configuration file (see the "server" directive below).
#no-resolv

# Specify IP address(es) of other DNS servers for queries not handled
# directly by consul. There is normally one 'server' entry set for every
# 'nameserver' parameter found in '/etc/resolv.conf'. See dnsmasq(8)'s
# 'server' configuration option for details.
#server=1.2.3.4
#server=208.67.222.222
#server=8.8.8.8

# Set the size of dnsmasq's cache. The default is 150 names. Setting the
# cache size to zero disables caching.
#cache-size=65536

» Unbound Setup

Unbound is typically configured via a unbound.conf or a series of files in the /etc/unbound/unbound.conf.d directory. In an Unbound configuration file (e.g. /etc/unbound/unbound.conf.d/consul.conf), add the following:

#Allow insecure queries to local resolvers
server:
  do-not-query-localhost: no
  domain-insecure: "consul"

#Add consul as a stub-zone
stub-zone:
  name: "consul"
  stub-addr: 127.0.0.1@8600

You may have to add the following line to the bottom of your /etc/unbound/unbound.conf file for the new configuration to be included:

include: "/etc/unbound/unbound.conf.d/*.conf"

» systemd-resolved Setup

systemd-resolved is typically configured with /etc/systemd/resolved.conf. To configure systemd-resolved to send queries for the consul domain to Consul, configure resolved.conf to contain the following:

DNS=127.0.0.1
Domains=~consul

The main limitation with this configuration is that the DNS field cannot contain ports. So for this to work either Consul must be configured to listen on port 53 instead of 8600 or you can use iptables to map port 53 to 8600. The following iptables commands are sufficient to do the port mapping.

[root@localhost ~]# iptables -t nat -A OUTPUT -d localhost -p udp -m udp --dport 53 -j REDIRECT --to-ports 8600
[root@localhost ~]# iptables -t nat -A OUTPUT -d localhost -p tcp -m tcp --dport 53 -j REDIRECT --to-ports 8600

Note: With this setup, PTR record queries will still be sent out to the other configured resolvers in addition to Consul.

» iptables Setup

On Linux systems that support it, incoming requests and requests to the local host can use iptables to forward ports on the same machine without a secondary service. Since Consul, by default, only resolves the .consul TLD, it is especially important to use the recursors option if you wish the iptables setup to resolve for other domains. The recursors should not include the local host as the redirects would just intercept the requests.

The iptables method is suited for situations where an external DNS service is already running in your infrastructure and is used as the recursor or if you want to use an existing DNS server as your query endpoint and forward requests for the consul domain to the Consul server. In both of those cases you may want to query the Consul server but not need the overhead of a separate service on the Consul host.

[root@localhost ~]# iptables -t nat -A PREROUTING -p udp -m udp --dport 53 -j REDIRECT --to-ports 8600
[root@localhost ~]# iptables -t nat -A PREROUTING -p tcp -m tcp --dport 53 -j REDIRECT --to-ports 8600
[root@localhost ~]# iptables -t nat -A OUTPUT -d localhost -p udp -m udp --dport 53 -j REDIRECT --to-ports 8600
[root@localhost ~]# iptables -t nat -A OUTPUT -d localhost -p tcp -m tcp --dport 53 -j REDIRECT --to-ports 8600

» macOS Setup

On macOS systems, you can use the macOS system resolver to point all .consul requests to consul. Just add a resolver entry in /etc/resolver/ to point at consul. documentation for this feature is available via: man5 resolver. To setup create a new file /etc/resolver/consul (you will need sudo/root access) and put in the file:

nameserver 127.0.0.1
port 8600

This is telling the macOS resolver daemon for all .consul TLD requests, ask 127.0.0.1 on port 8600.

» Testing

First, perform a DNS query against Consul directly to be sure that the record exists:

[root@localhost ~]# dig @localhost -p 8600 primary.redis.service.dc-1.consul. A

; <<>> DiG 9.8.2rc1-RedHat-9.8.2-0.23.rc1.32.amzn1 <<>> @localhost primary.redis.service.dc-1.consul. A
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 11536
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0

;; QUESTION SECTION:
;primary.redis.service.dc-1.consul. IN A

;; ANSWER SECTION:
primary.redis.service.dc-1.consul. 0 IN A 172.31.3.234

;; Query time: 4 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Wed Apr  9 17:36:12 2014
;; MSG SIZE  rcvd: 76

Then run the same query against your BIND instance and make sure you get a valid result:

[root@localhost ~]# dig @localhost -p 53 primary.redis.service.dc-1.consul. A

; <<>> DiG 9.8.2rc1-RedHat-9.8.2-0.23.rc1.32.amzn1 <<>> @localhost primary.redis.service.dc-1.consul. A
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 11536
;; flags: qr rd ra; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0

;; QUESTION SECTION:
;primary.redis.service.dc-1.consul. IN A

;; ANSWER SECTION:
primary.redis.service.dc-1.consul. 0 IN A 172.31.3.234

;; Query time: 4 msec
;; SERVER: 127.0.0.1#53(127.0.0.1)
;; WHEN: Wed Apr  9 17:36:12 2014
;; MSG SIZE  rcvd: 76

If desired, verify reverse DNS using the same methodology:

[root@localhost ~]# dig @127.0.0.1 -p 8600 133.139.16.172.in-addr.arpa. PTR

; <<>> DiG 9.10.3-P3 <<>> @127.0.0.1 -p 8600 133.139.16.172.in-addr.arpa. PTR
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 3713
;; flags: qr aa rd; QUERY: 1, ANSWER: 1, AUTHORITY: 0, ADDITIONAL: 0
;; WARNING: recursion requested but not available

;; QUESTION SECTION:
;133.139.16.172.in-addr.arpa.   IN  PTR

;; ANSWER SECTION:
133.139.16.172.in-addr.arpa. 0  IN  PTR consul1.node.dc1.consul.

;; Query time: 3 msec
;; SERVER: 127.0.0.1#8600(127.0.0.1)
;; WHEN: Sun Jan 31 04:25:39 UTC 2016
;; MSG SIZE  rcvd: 109
[root@localhost ~]# dig @127.0.0.1 +short -x 172.16.139.133
consul1.node.dc1.consul.

» Troubleshooting

If you don't get an answer from your DNS server (e.g. BIND, Dnsmasq) but you do get an answer from Consul, your best bet is to turn on your DNS server's query log to see what's happening.

For BIND:

[root@localhost ~]# rndc querylog
[root@localhost ~]# tail -f /var/log/messages

The log may show errors like this:

error (no valid RRSIG) resolving
error (no valid DS) resolving

This indicates that DNSSEC is not disabled properly.

If you see errors about network connections, verify that there are no firewall or routing problems between the servers running BIND and Consul.

For Dnsmasq, see the log-queries configuration option and the USR1 signal.