coredns/plugin/etcd/README.md

# etcd

## Name

*etcd* - enables SkyDNS service discovery from etcd.

## Description

The *etcd* plugin implements the (older) SkyDNS service discovery service. It is *not* suitable as
a generic DNS zone data plugin. Only a subset of DNS record types are implemented, and subdomains
and delegations are not handled at all. The plugin will also recursively descend the tree and return
all records found, see "Special Behavior" below for details.

The data in the etcd instance has to be encoded as
a [message](https://github.com/skynetservices/skydns/blob/2fcff74cdc9f9a7dd64189a447ef27ac354b725f/msg/service.go#L26)
like [SkyDNS](https://github.com/skynetservices/skydns). It works just like SkyDNS.

The *etcd* plugin makes extensive use of the *forward* plugin to forward and query other servers in the
network - if that plugin has been enabled as well.

## Syntax

~~~
etcd [ZONES...]
~~~

* **ZONES** zones *etcd* should be authoritative for.

The path will default to `/skydns` the local etcd3 proxy (http://localhost:2379). If no zones are
specified the block's zone will be used as the zone.


~~~
etcd [ZONES...] {
    fallthrough [ZONES...]
    path PATH
    endpoint ENDPOINT...
    credentials USERNAME PASSWORD
    tls CERT KEY CACERT
}
~~~

* `fallthrough` If zone matches but no record can be generated, pass request to the next plugin.
  If **[ZONES...]** is omitted, then fallthrough happens for all zones for which the plugin
  is authoritative. If specific zones are listed (for example `in-addr.arpa` and `ip6.arpa`), then only
  queries for those zones will be subject to fallthrough.
* **PATH** the path inside etcd. Defaults to "/skydns".
* **ENDPOINT** the etcd endpoints. Defaults to "http://localhost:2379".
* `credentials` is used to set the **USERNAME** and **PASSWORD** for accessing the etcd cluster.
* `tls` followed by:

    * no arguments, if the server certificate is signed by a system-installed CA and no client cert is needed
    * a single argument that is the CA PEM file, if the server cert is not signed by a system CA and no client cert is needed
    * two arguments - path to cert PEM file, the path to private key PEM file - if the server certificate is signed by a system-installed CA and a client certificate is needed
    * three arguments - path to cert PEM file, path to client private key PEM file, path to CA PEM
      file - if the server certificate is not signed by a system-installed CA and client certificate
      is needed.

## Special Behaviour

The *etcd* plugin leverages directory structure to look for related entries. For example
an entry `/skydns/test/skydns/mx` would have entries like `/skydns/test/skydns/mx/a`,
`/skydns/test/skydns/mx/b` and so on. Similarly a directory `/skydns/test/skydns/mx1` will have all
`mx1` entries. Note this plugin will search through the entire (sub)tree for records. In case of the
first example, a query for `mx.skydns.test` will return both the contents of the `a` and `b` records.
If the directory extends deeper those records are returned as well.

With etcd3, support for [hierarchical keys are
dropped](https://coreos.com/etcd/docs/latest/learning/api.html). This means there are no directories
but only flat keys with prefixes in etcd3. To accommodate lookups, the *etcd* plugin now does a lookup
on prefix `/skydns/test/skydns/mx/` to search for entries like `/skydns/test/skydns/mx/a` etc, and
if there is nothing found on `/skydns/test/skydns/mx/`, it looks for `/skydns/test/skydns/mx` to
find entries like `/skydns/test/skydns/mx1`.

This causes two lookups from CoreDNS to etcd in certain cases.

## Examples

This is the default SkyDNS setup, with everything specified in full:

~~~ corefile
skydns.local {
    etcd {
        path /skydns
        endpoint http://localhost:2379
    }
    prometheus
    cache
    loadbalance
}

. {
    forward . 8.8.8.8:53 8.8.4.4:53
    cache
}
~~~

Or a setup where we use `/etc/resolv.conf` as the basis for the proxy and the upstream
when resolving external pointing CNAMEs.

~~~ corefile
skydns.local {
    etcd {
        path /skydns
    }
    cache
}

. {
    forward . /etc/resolv.conf
    cache
}
~~~

Multiple endpoints are supported as well.

~~~
etcd skydns.local {
    endpoint http://localhost:2379 http://localhost:4001
...
~~~
Before getting started with these examples, please setup `etcdctl` (with `etcdv3` API) as explained
[here](https://coreos.com/etcd/docs/latest/dev-guide/interacting_v3.html). This will help you to put
sample keys in your etcd server.

If you prefer, you can use `curl` to populate the `etcd` server, but with `curl` the
endpoint URL depends on the version of `etcd`. For instance, `etcd v3.2` or before uses only
[CLIENT-URL]/v3alpha/* while `etcd v3.5` or later uses [CLIENT-URL]/v3/* . Also, Key and Value must
be base64 encoded in the JSON payload. With `etcdctl` these details are automatically taken care
of. You can check [this document](https://github.com/coreos/etcd/blob/master/Documentation/dev-guide/api_grpc_gateway.md#notes)
for details.

### Reverse zones

Reverse zones are supported. You need to make CoreDNS aware of the fact that you are also
authoritative for the reverse. For instance if you want to add the reverse for 10.0.0.0/24, you'll
need to add the zone `0.0.10.in-addr.arpa` to the list of zones. Showing a snippet of a Corefile:

~~~
etcd skydns.local 10.0.0.0/24 {
...
~~~

Next you'll need to populate the zone with reverse records, here we add a reverse for
10.0.0.127 pointing to reverse.skydns.local.

~~~
% etcdctl put /skydns/arpa/in-addr/10/0/0/127 '{"host":"reverse.skydns.local."}'
~~~

Querying with dig:

~~~ sh
% dig @localhost -x 10.0.0.127 +short
reverse.skydns.local.
~~~

### Zone name as A record

The zone name itself can be used as an `A` record. This behavior can be achieved by writing special
entries to the ETCD path of your zone. If your zone is named `skydns.local` for example, you can
create an `A` record for this zone as follows:

~~~
% etcdctl put /skydns/local/skydns/ '{"host":"1.1.1.1","ttl":60}'
~~~

If you query the zone name itself, you will receive the created `A` record:

~~~ sh
% dig +short skydns.local @localhost
1.1.1.1
~~~

If you would like to use DNS RR for the zone name, you can set the following:
~~~
% etcdctl put /skydns/local/skydns/x1 '{"host":"1.1.1.1","ttl":60}'
% etcdctl put /skydns/local/skydns/x2 '{"host":"1.1.1.2","ttl":60}'
~~~

If you query the zone name now, you will get the following response:

~~~ sh
% dig +short skydns.local @localhost
1.1.1.1
1.1.1.2
~~~

### Zone name as AAAA record

If you would like to use `AAAA` records for the zone name too, you can set the following:
~~~
% etcdctl put /skydns/local/skydns/x3 '{"host":"2003::8:1","ttl":60}'
% etcdctl put /skydns/local/skydns/x4 '{"host":"2003::8:2","ttl":60}'
~~~

If you query the zone name for `AAAA` now, you will get the following response:
~~~ sh
% dig +short skydns.local AAAA @localhost
2003::8:1
2003::8:2
~~~

### SRV record

If you would like to use `SRV` records, you can set the following:
~~~
% etcdctl put /skydns/local/skydns/x5 '{"host":"skydns-local.server","ttl":60,"priority":10,"port":8080}'
~~~
Please notice that the key `host` is the `target` in `SRV`, so it should be a domain name.

If you query the zone name for `SRV` now, you will get the following response:

~~~ sh
% dig +short skydns.local SRV @localhost
10 100 8080 skydns-local.server.
~~~

### TXT record

If you would like to use `TXT` records, you can set the following:
~~~
% etcdctl put /skydns/local/skydns/x6 '{"ttl":60,"text":"this is a random text message."}'
% etcdctl put /skydns/local/skydns/x7 '{"ttl":60,"text":"this is a another random text message."}'
~~~

If you query the zone name for `TXT` now, you will get the following response:
~~~ sh
% dig +short skydns.local TXT @localhost
"this is a random text message."
"this is a another random text message."
~~~

## See Also

If you want to `round robin` A and AAAA responses look at the *loadbalance* plugin.