#nodepp
An EPP implementation in node.js
##Description
This is a service for communicating with registries over EPP. It takes datastructures in JSON, converts them to XML, sends them to the registry, and then does the whole thing in reverse with the response.
There are two separate server scripts:
lib/node-epp-server.js
is designed to function as a RESTful interface where you can POST and receive json requests.lib/rabbit-epp.js
, runs as an RPC server that accepts requests via RabbitMQ.
- Clone the repository:
git clone https://github.com/heytrav/nodepp.git
. cd nodepp
npm install
to install node dependencies.source nodepp.rc
to include./node_modules/.bin
in the$PATH
. This is necessary for testing or if you plan on running it as a daemon.
In config/epp-config-example.json
you'll find something that looks like
this:
{
"registries": {
"registry-test1": {
// registry1 data
},
"registry-test2": {
// registry2
}
},
"rabbitmq": {
"connection": {
// RabbitMQ connection
}
},
"whitelisted_ips": []
}
For your development and production environments, I recommend copying this
file to config/epp-config-devel.json
and
config/epp-config-production.json
, respectively, and modifying each to
fit your needs. You will need to add your own login/password as well as the
paths to any SSL certificates and keys. You can of course replace the keys
registry-test1
, registry-test2
, etc. with something more descriptive of the registry.
When you've got the config setup the way you like it, symlink this to
config/epp-config.json
to run the application.
ln -s <path to app>/config/epp-config-devel.json <path to app>/config/epp-config.json
You can add as many registries as you like. You may even need to add the same registry multiple times with different logins, etc. This is practical for testing if you need to simulate transferring domains between two separate registrars.
The whitelisted_ips
tells the REST application to only accept certain hosts.
npm test
Note that a number of tests are currently set to skip automatically. These require a running RabbitMQ instance and that you have set up the configuration accordingly. They also assume that you have an online testing environment (OTE) account with some registry or registrar.
The REST app is based on express.js and listens for POST requests on port 3000. The general URL scheme is as follows:
http://<host>:3000/command/<registry>/<command>
So to run a checkDomain for registry1 on a local instance of the server, POST your request to:
http://localhost:3000/command/registry1/checkDomain
You can start it as follows:
node lib/node-epp-server.js -r registry1 -r registry2
The -r
option specifies which registries the script should log into and can be used multiple times. These should correspond to the registries
listed in the configuration file. In this case, the client will log into into registry1 and registry2.
Alternatively you can start it as a daemon:
foreverd start -o nodepp-stout.log -e nodepp-sterr.log lib/node-epp-server.js \
-r registry-test1 -r registry-test2 -r registry-test3
This tells it to open connections to three different registries.
To stop the service:
foreverd stop lib/node-epp-server.js
You can test the script by posting JSON requests to the server instance. I recommend using the program Postman which can be installed in Chrome/Firefox as an extension. However, you can also use curl or the scripting language of your choice. I put an example of this down below.
This is a RPC service that listens for connections on RabbitMQ. Have a look at the rabbitmq
section of the configuration file if you would like to run this. To run the application:
node lib/rabbit-epp.js -r registry-test1
Or as a daemon:
foreverd start -o nodepp-stout.log -e nodepp-sterr.log lib/rabbit-epp.js \
-r registry-test1 -r registry-test2 -r registry-test3
To stop it:
foreverd stop lib/rabbit-epp.js
Note that for all commands documented below, the datastructure that is sent via RabbitMQ needs to be modified as follows:
{
"command": "<command name>",
"data": <request data>
}
I wrote some scripts for interacting with the RabbitMQ service in my epp-reg project. Feel free to use those as you like.
{"id": "P-12345xyz"}
or
{"contact": "P-12345xyz"}
{"id": "P-12345xyz"}
or
{"contact": "P-12345xyz"}
{
"id": "my-id-1234",
"voice": "+1.9405551234",
"fax": "+1.9405551233",
"email": "[email protected]",
"authInfo": {
"pw": "xyz123"
},
"postalInfo": [{
"name": "John Doe",
"org": "Example Ltd",
"type": "int",
"addr": [{
"street": ["742 Evergreen Terrace", "Apt b"],
"city": "Springfield",
"sp": "OR",
"pc": "97801",
"cc": "US"
}]
}]
}
Some registries set the id
by default. In such cases it's common to use
auto
. The value for type
may also vary for different
registries. Some require loc
and some require int
. EPP allows for
up to 2 postaInfo entries, however I've never seen a registry that accepts
more than 1. For that reason, you can just specify it as a single object:
"postalInfo": {
"name": "John Doe",
"org": "Example Ltd",
"type": "int",
"addr": [{
"street": ["742 Evergreen Terrace", "Apt b"],
"city": "Springfield",
"sp": "OR",
"pc": "97801",
"cc": "US"
}]
}
It will be passed to the registry in the appropriate format. The same applies to the addr field (in postalInfo), which can also be specified as an Array or single object.
"addr": {
"street": ["742 Evergreen Terrace", "Apt b"],
"city": "Springfield",
"sp": "OR",
"pc": "97801",
"cc": "US"
}
{
id: "p-12345",
add: ['clientDeleteProhibited'],
rem: ['clientTransferProhibited'],
chg: {
"postalInfo": [{
"name": "John Doe",
"org": "Example Ltd",
"type": "loc",
"addr": [{
"street": ["742 Evergreen Terrace", "Apt b"],
"city": "Eugene",
"sp": "OR",
"pc": "97801",
"cc": "US"
}]
}],
"voice": "+1.9405551234",
"fax": "+1.9405551233",
"email": "[email protected]",
"authInfo": {
"pw": "xyz123"
},
"disclose": {
"flag": 0,
"disclosing": ["voice", "email"]
}
}
}
The following are equivalent:
{"domain": "something.com"}
or
{"name": "something.com"}
It is possible to check more than one domain at a time.
{"domain": ["test-domain.com", "test-domain2.com", "test-domain3.com"]}
{"domain": "something.com"}
In case you are wondering if you can send multiple domains like in checkDomain, the answer is no. That's not possible in EPP. The result that you will get back in one infoDomain will be complicated enough.
{
"name": "myreg-test-101-domain.com",
"period": {
"unit": "y",
"value": 1
},
"ns":[
"ns1.hexonet.net",
"ns2.hexonet.net"
],
"registrant": "my-id-1234",
"contact": [
{ "admin": "my-id-1235" },
{ "tech": "my-id-1236" },
{"billing": "my-id-1236"}
],
"authInfo": {
"pw": "Axri3k.XXjp"
}
}
See comments below regarding alternative formats for ns, period, and authInfo fields.
{"domain": "something.com"}
{
"domain": "something.com",
"curExpDate": "2016-04-03",
"period": {
"unit": 'y',
"value": 1
}
}
period is optional and will default to 1 year.
{
"name": "test-domain.com",
"op": "request",
"period": 1,
"authInfo": {
"roid": "P-12345", // optional
"pw": "2fooBAR"
}
}
Valid values for op
are approve, cancel, query, reject, and
request. There uses are:
-
Requesting side
- request to request a transfer.
- cancel to cancel a transfer.
- query to find out if a transfer is pending (although we should get info via polling)
-
Domain holder side
- approve to approve a transfer request from another registrar.
- reject to reject a transfer request from another registrar.
{
"name": "test-domain.com",
"add": {
"ns": ["ns3.test.com", "ns4.whatever.com"],
"contact": [{
"admin": "P-9876"
},
{
"billing": "PX143"
}],
"status": ["clientUpdateProhibited", {
"s": "clientHold",
"lang": "en",
"value": "Payment Overdue"
}]
},
"rem": {
"ns": [{
"host": "ns1.test-domain.com",
"addr": {
"type": "v4",
"ip": "192.68.2.132"
}
}],
"contact": [{
"billing": "PX147"
}],
"status": ["clientTransferProhibited", {
"s": "clientWhatever",
"lang": "en",
"value": "Payment Overdue"
}]
},
"chg": {
"registrant": "P-49023",
"authInfo": {
"pw": "TestPass2"
}
}
}
This is a very complicated example but at least shows what is possible in an
updateDomain. At least 1 of add
, rem
, or chg
is required.
The chg
field, if provided, must contain either a registrant
and/or an authInfo
. add
and rem
elements, if provided, must
contain any one or more ns
, contact
, or status
fields.
{
"name": "ns1.host.com",
"addr": ["23.84.43.123", {
"ip": "22.4.22.5"
},
{
"ip": "::F3:34::BA:",
"type": "v6"
}]
}
{
"name": "ns1.host.com",
"chg": {
"name": "ns2.host.com",
},
"add": {
"addr": {
"ip": "::F3:34::BA:",
"type": "v6"
},
"status": ["clientUpdateProhibited"]
},
"rem": {
"addr": ["23.84.43.123", {
"ip": "22.4.22.5"
}],
"status": ["clientTransferProhibited", "sneezeAchoo"]
}
}
Some of the required datastructures might seem a bit weird. EPP has a fairly
complex grammar that is probably intended to make granular control of domain
related entities possible. There are no flat datastructures and some things
must be specified explicitly that would be assumed in systems that use
key=value
APIs. For example, to remove nameservers from a domain, it is
necessary to remove them explicitly. Simply updating domain with the new
nameservers will not work. The same goes for contact objects.
In createDomain and updateDomain I've tried to account for 2 different types of host objects. In the simplest version you can just pass an array of strings:
["ns1.host.com", "ns2.host.tld"]
In cases where IP addresses are required, the following variants can be used:
[{host: "ns1.host.com", addr: "62.47.23.1"}]
[{host: "ns2.host.com", addr:[ "62.47.23.1", {ip: "53.23.1.3"} ]}]
[{host: "ns3.host.com", addr:[ {ip: "::F3:E2:23:::", type: "v6"}, {ip:"47.23.43.1", type: "v4"} ]}]
type
is v4
by default. You'll have to specify v6
explicitly for IPv6
addresses.
It's up to you to know which cases glue records are required. This implementation has no way to know that.
createContact, createDomain, transferDomain and updateDomain accept an
authInfo
parameter.
Following are equivalent:
authInfo: "te2tP422t"
or
authInfo: {
pw: "te2tP422t"
}
In some cases you may need to supply a roid
in addition to the
authInfo
. This is used to identify the registrant or contact object if
and only if the given authInfo is associated with a registrant or contact
object, and not the domain object itself.
authInfo: {
pw: "te2tP422t",
roid: "P-1234"
}
The period
argument in createDomain, renewDomain and transferDomain can be specified as follows:
1 year registration
period: 1
24 month registration
period: {
unit: "m",
value: 24
}
The default unit is y for year and default period is 1.
A transactionId
is optional. It can be added at the top level of the JSON data
structure. By default it will be set to myreg-<epoch timestamp>
.
You can optionally add an extension
property to some commands. This
varies from registry to registry like everything else. A good example is when
adding DNSSEC
data to a createDomain:
{
"name": "myreg-test-101-domain.com",
"period": {
"unit": "y",
"value": 1
},
"ns":["ns1.hexonet.net","ns2.hexonet.net"],
"registrant": "my-id-1234",
"contact": [
{ "admin": "my-id-1235" },
{ "tech": "my-id-1236" },
{"billing": "my-id-1236"}
],
"authInfo": {
"pw": "Axri3k.XXjp"
},
"extension": {
"DNSSEC": {
"maxSigLife": 604800,
"dsData": {
"keyTag": 12345,
"alg": 3,
"digestType": 1,
"digest": "49FD46E6C4B45C55D4AC",
"keyData": {
"flags": 257,
"protocol": 3,
"alg": 1,
"pubKey": "AQPJ////4Q=="
}
}
}
}
}
I've implemented DNSSEC
EPP generation for create and update.
Following are some variations that you can send (I'm leaving out the standard part of the EPP request):
Create a domain with the dsData
interface:
"extension": {
"DNSSEC": {
"maxSigLife": 604800,
"dsData": {
"keyTag": 12345,
"alg": 3,
"digestType": 1,
"digest": "49FD46E6C4B45C55D4AC"
}
}
}
with the keyData
interface:
"extension": {
"DNSSEC": {
"keyData":{
"flags": 257,
"protocol": 3,
"alg": 1,
"pubKey": "AQPJ////4Q=="
}
}
}
with the keyData
in the dsData
element:
"extension": {
"DNSSEC": {
"maxSigLife": 604800,
"dsData": {
"keyTag": 12345,
"alg": 3,
"digestType": 1,
"digest": "49FD46E6C4B45C55D4AC",
"keyData":{
"flags": 257,
"protocol": 3,
"alg": 1,
"pubKey": "AQPJ////4Q=="
}
}
}
}
Add a dsData
key and remove a keyData
key and change the
maxSigLife
of the key
"extension": {
"DNSSEC": {
"add": {
"dsData": {
"keyTag": 12345,
"alg": 3,
"digestType": 1,
"digest": "49FD46E6C4B45C55D4AC"
}
},
"rem": {
"keyData": {
"flags": 257,
"protocol": 3,
"alg": 1,
"pubKey": "AQPJ////4Q=="
}
},
"chg": {
"maxSigLife": 604800
}
}
}
Remove all existing key info and replace it with something new:
"extension": {
"DNSSEC": {
"rem": {
"all": true
},
"add": {
"dsData": {
"keyTag": 12345,
"alg": 3,
"digestType": 1,
"digest": "49FD46E6C4B45C55D4AC"
}
}
}
}
Post the following to http://localhost:3000/command/hexonet/checkDomain
prompt$ time curl -H "Content-Type: application/json" \
-d '{"domain": "test-domain.com"}' \
http://localhost:3000/command/<registry>/checkDomain
Note I just put time
in there to show what the execution time is.
YMMV
You should get the following response (or something similar):
"result":{
"code":1000,"msg":"Command completed successfully"},
"resData":{
"domain:chkData": {
"xmlns:domain":"urn:ietf:params:xml:ns:domain-1.0",
"xsi:schemaLocation":"urn:ietf:params:xml:ns:domain-1.0 domain-1.0.xsd",
"domain:cd":{
"domain:name":{
"avail":0,
"$t":"test-domain.com"
},
"domain:reason":"Domain exists"
}
}
},
I plan to get rid of some of the EPP cruft in the near future.