Route 53 DynDNS API

This project is no longer active

Rather than upgrade this to Python3, I've moved to using ddns-route53 running in a kube pod in my home network.

Overview

This project implements a DynDNS v2 API compatible front end for the AWS Route 53 service. This allows most consumer routers and any other DynDNS compatible clients to update A records in Route 53 hosted domains.

This is implemented as an AWS Lambda function fronted by the AWS API Gateway service. With updates once an hour, this service can be run entirely within the free tier for these services.

Prerequisites

Amazon Web Services Account
Domain hosted with AWS Route 53
Client that speaks the DynDNS v2 API (most common)
Terraform (for install, tested with 0.9.2)

Setup

Step 1: Terraform Credentials

Terraform needs credentials in order to be able to create AWS resources. The simplest way to do this is to set AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY and AWS_DEFAULT_REGION environment variables.

Alternatively, if you have the AWS CLI tools installed and configured, you can get away with setting just AWS_DEFAULT_REGION and Terraform will use the credentials you've configured for use with them.

More details on other options can be found in the Terraform documentation for AWS Authentication.

Step 2: Configuration

You will need to configure at a minimum the username and password you want clients to use for authentication. The simplest way to do this is to copy terraform.tfvars.example to terraform.tfvars then edit the file.

The password key is the only value in this file that must be changed.

The username key should be self-explanatory.

The default TTL on records when created or updated is 5 minutes (300 seconds). If you wish to change that value then uncomment the ttl key in the config file and set it to the number of seconds you want the TTL to be set to.

Step 3: Deploying

There are two ways to deploy this to AWS:

Quick And Easy

Type make. This will build the zip files for deployment and run terraform apply. If it works, the last line of output should contain the url for your API endpoint like this:

Outputs:

url = https://24ufed9k37.execute-api.us-east-2.amazonaws.com/nic

Step By Step

Run make zip. This will build the deployment packages for the main function and the custom authorizer.
Run terraform plan. This will validate that your AWS credentials are working properly, evaluate existing state and describe what it will do when you deploy.
Run terraform apply. This will deploy the functions, API gateway and all other supporting resources. If it works, the last line of output should contain the url for your API endpoint like this:

Outputs:

url = https://24ufed9k37.execute-api.us-east-2.amazonaws.com/nic

Step 4: Configuring your client

Documenting the configuration for every possible client is impossible. For the most part this should work with any client that supports the DynDNS v2 API that allows you to specify a different server name. Generally you want to configure just like any other DynDNS v2 API service, but make sure you specify the server as the value returned in the last line of the terraform apply output.

Here is an example config that works with ddclient:

daemon=1m
syslog=yes
ssl=yes
use=if, if=eth0

server=24ufed9k37.execute-api.us-east-2.amazonaws.com,protocol=dyndns2
login=route53-ddns
password=<password here>
your.hostname.com

Example configuration for a Ubiquiti EdgeRouter:

service {
    dns {
        dynamic {
            interface eth0 {
                service dyndns {
                    host-name your.hostname.com
                    login route53-ddns
                    password <password here>
                    protocol dyndns2
                    server 24ufed9k37.execute-api.us-east-2.amazonaws.com
                }
            }
        }
    }

Updating

If you update this code, the simplest way to update is just to run make again. That will update deployment archives if needed and run terraform apply for you automatically.

Debugging

If you encounter errors being sent back to your client, there are a few things to try.

First, check CloudWatch. You should have a pair of log groups named /aws/lambda/route53-ddns and /aws/lambda/route53-ddns-authorizer. The authorizer is used only for checking the username and password. If it allows access, then the main function will be called. It can be helpful to manually force an update and look to see if a new log event shows up in both the authorizer and the main function. If no log even shows up in either, the error may be in the API Gateway itself. Unfortunately this can be very difficult to troubleshoot. In theory you can enable CloudWatch log groups for API gateway, but I've had limited success with this.

If you are getting logs for requests, but still can't figure out what is going on, you should try turning on debug logging via the Terraform debug key. This will enable debug level logging. This can be a bit overwhelming, since it includes all the boto3 logs, but generally has worked well for me.

You may also want to manually force an update via curl. In this case the output from the function or the error message returned may be helpful. For example:

curl -v -G -d hostname=<hostname> https://route53-ddns:<password here>@<api endpoint>/nic/update

Notes

The Terraform plans don't currently support deploying to a endpoint in a custom domain. It would be nice to add this in the future.
The normal DynDNS v2 API endpoint path is /nic/update, but the function isn't particular about the path. Anything under /nic will work, and you can configure a custom stage name via Terraform if you want a different base path.

Known Issues

Sometimes you may see an error like below when initially running terraform apply:

Error applying plan:

1 error(s) occurred:

* aws_api_gateway_deployment.deployment: 1 error(s) occurred:

* aws_api_gateway_deployment.deployment: Error creating API Gateway Deployment: BadRequestException: No integration defined for method
        status code: 400, request id: e32e0d75-1b02-11e7-ac81-e98a6fba3c24

This appears to be caused by a race condition during creation where the integration is not yet fully created in AWS when the deployment is triggered. Running terraform apply again will succeessfully execute the plan.

When running Terraform in plan or apply mode, it will always report that the aws_api_gateway_deployment.deployment resource is changed. This is intentional and is intended to force a deployment on every deploy. This is needed current due to Terraform issue #6613
Due to API Gateway limitations, Basic Authentication isn't implemented completely correctly. This is because it's not possible for a custom authenticator to control the headers sent, so the WWW-Authenticate header cannot be sent back. This isn't needed for DDClient or other clients that send the Authorization in the initial response. Please let me know if you encounter any client that has problems with it.
When a change is accepted, it is queued in Route 53 for update, but will still be pending for up to a few minutes. Because of this when a change is submitted via the API it will always return with the "good" response instead of "nochg", even if the IP associated with the record hasn't changed.

Running Tests

Tests are setup to be run via tox. There is one test environment for unit tests and another for PEP8 tests. Once tox is installed just run tox to run both.

Author

Clayton O'Neill clayton@oneill.net

claytono / route53-ddns