This tutorial guides you through the steps for configuring Marathon-LB to run a containerized DC/OS service with external load balancing for a website. After completing this tutorial, you will have hands-on practice configuring load balancing for a sample application.

In this tutorial, you will:

• Download a Docker image that contains NGINX as a sample app.
• Prepare load balancing for the NGINX application on the dcos.io website.
• Install Marathon-LB as the edge load balancer for the service.
• Run Marathon-LB on a public-facing node to route inbound network traffic.

Before you begin

• You must have a DC/OS cluster with a bootstrap node, at least one master node, at least one private agent node, and at least one public agent node.
• You must have an account with access to the DC/OS web-based administrative console or DC/OS command-line interface.
• You must have Marathon-LB installed.

Download the app definition and container

This tutorial uses a sample app definition file that you can download from the dcos-website GitHub repository.

1. Copy dcos-website/dcos-website.json from the dcos-website GitHub repository.

2. Go to the mesosphere/dcos-website Docker repository and copy the latest image tag.

For example, after clicking Tags, you might see an identifier similar to this: cff383e4f5a51bf04e2d0177c5023e7cebcab3cc

Modify the app image tag and public IP address

1. Open the dcos-website.json app definition file you downloaded from the repository.

2. Add the image-tag in the docker:image field with the Docker image tag.

   For example:

   {
     "id": "dcos-website",
     "container": {
       "type": "DOCKER",
       "docker": {
         "image": "mesosphere/dcos-website:cff383e4f5a51bf04e2d0177c5023e7cebcab3cc",
         "network": "BRIDGE",
         "portMappings": [
           { "hostPort": 0, "containerPort": 80, "servicePort": 10004 }
         ]
       }
     },
   

3. Identify the public-facing IP address for your public agent node.

   For information about how to find your public agent IP address, see Finding a public agent IP.

4. In the labels field, add an entry for HAPROXY_0_VHOST and assign it the value of your public agent IP.

   For example, if the public agent node IP address is 64.172.103.2, you might add lines similar to the following:

   "labels":{
           "HAPROXY_DEPLOYMENT_GROUP":"dcos-website",
           "HAPROXY_DEPLOYMENT_ALT_PORT":"10005",
           "HAPROXY_GROUP":"external",
           "HAPROXY_0_REDIRECT_TO_HTTPS":"true",
           "HAPROXY_0_VHOST": "64.172.103.2"
         }
   

   Be sure to remove the leading http:// and the trailing slash () from the IP address, and to add a comma after the preceding field.

   The complete JSON app definition file should resemble the following:

     {
       "id": "dcos-website",
       "container": {
         "type": "DOCKER",
         "portMappings": [
           { "hostPort": 0, "containerPort": 80, "servicePort": 10004 }
         ],
         "docker": {
           "image": "mesosphere/dcos-website:cff383e4f5a51bf04e2d0177c5023e7cebcab3cc"
         }
       },
       "instances": 3,
       "cpus": 0.25,
       "mem": 100,
       "network": "BRIDGE",
       "healthChecks": [{
           "protocol": "HTTP",
           "path": "/",
           "portIndex": 0,
           "timeoutSeconds": 2,
           "gracePeriodSeconds": 15,
           "intervalSeconds": 3,
           "maxConsecutiveFailures": 2
       }],
       "labels":{
         "HAPROXY_DEPLOYMENT_GROUP":"dcos-website",
         "HAPROXY_DEPLOYMENT_ALT_PORT":"10005",
         "HAPROXY_GROUP":"external",
         "HAPROXY_0_REDIRECT_TO_HTTPS":"true",
         "HAPROXY_0_VHOST": "64.172.103.2"
       }
     }
   

Only apps with the label HAPROXY_GROUP=external will be exposed using this Marathon-LB configuration.

Add the load-balanced application and check its status

1. Run the service from the DC/OS CLI using the following command:

   dcos marathon app add dcos-website.json
   

2. Open the DC/OS web-based console URL in a browser, then click Services to verify that your application is deployed and running.

   

3. In the web browser, navigate to the IP address for your public agent node to verify the site you have deployed is running.