This document explains how to install DC/OS 1.10 using the Azure Resource Manager templates.

Tip: To get support on Azure Marketplace-related questions, join the Azure Marketplace Slack community.

Important: Upgrades are not supported with this installation method.

System requirements


To use all of the services offered in DC/OS, you should choose at least five Mesos Agents using Standard_D2 Virtual Machines, which is the default size in the DC/OS Azure Marketplace offering.

Selecting smaller-sized VMs is not recommended, and selecting fewer VMs will likely cause certain resource-intensive services such as distributed datastores not to work properly (from installation issues to operational limitations).


You will need an active Azure subscription to install DC/OS via the Azure Marketplace.

Also, to access nodes in the DC/OS cluster you will need ssh installed and configured.

Install DC/OS

Step 1: Deploying the template

To install DC/OS 1.10 on Azure, use the Azure Resource Manager templates provided.

Some notes of the template configuration is below,

  • Choose East US as the Location, because some resources of the template may not available in other location.
  • Set Oauth Enabled to true if you want to sign in the DC/OS Dashboard through OAuth.
  • Fill up the Agent Endpoint DNS Name Prefix and Master Endpoint DNS Name Prefix.
  • Enter your Ssh RSA Public Key.

Step 2: Accessing DC/OS

First, look up MASTERFQDN in the outputs of the deployment. To find that, click on the link under Last deployment (which is 4/15/2016 (Succeeded) here) and you should see this:

Deployment history

Click on the latest deployment and copy the value of MASTERFQDN in the Outputs section:

Deployment output

Use the value of MASTERFQDN you found in the Outputs section in the previous step, and we will use it in the following step.

Because of security considerations, you can not visit the DC/OS Dashboard in Azure directly by default. Here, we provide two ways to work around. Please find your case below,

Case 1:

In order to visit the the DC/OS Dashboard, we will need to access the TCP port 80 or 443 of the master node. You can add an inbound security rule and an inbound NAT rule.

Find the network security group resource of the master node,

Resource - Master Node Network Security Group

Click on the “Inbound security rules” tab on the left side,

Inbound Security Rules

Add an inbound security rule.

Add Inbound Security Rules

Find the load balancer resource of the master node,

Resource - Master Node Load balancer

Click on the “Inbound NAT rules” tab on the left side,

Inbound NAT Rules

Add an inbound NAT rule.

Add Inbound NAT Rules

Now you can visit http://$MASTERFQDN and view the DC/OS Dashboard.

Case 2: Using ssh tunnel

In this case, we need to setup a ssh tunnel, to forward TCP port 80 of the master node on the azure to the 8000 port of your local machine.

Use the value of MASTERFQDN you found in the previous step and paste it in the following command:

ssh [email protected]$MASTERFQDN -L 8000:localhost:80

For example, in my case:

ssh [email protected] -L 8000:localhost:80

Now you can visit http://localhost:8000 on your local machine and view the DC/OS Dashboard.

DC/OS dashboard


Some caveats around SSH access:

  • For connections to http://localhost:8000 to work, the SSH command must be run on your local machine, and not inside a Virtual Machine.
  • In the example above, port 8000 is assumed to be available on your local machine.
  • The SSH commands shown only work on Mac or Linux. For Windows use Putty with a similar port-forwarding configuration, see also How to Use SSH with Windows on Azure.
  • If you want to learn more about SSH key generation check out this GitHub tutorial.

The DC/OS UI will not show the correct IP address or CLI install commands when connected by using an SSH tunnel.


The following commands can be used to run the DC/OS CLI directly on the master node:

# Connect to master node with ssh
ssh [email protected]$MASTERFQDN

# Install CLI on the master node and configure with http://localhost
curl -o dcos &&
sudo mv dcos /usr/local/bin &&
sudo chmod +x /usr/local/bin/dcos &&
dcos cluster setup http://localhost &&

# Now you can use the DC/OS CLI:
dcos package search

Tear down the DC/OS cluster

If you’ve created a new resource group in the deployment step, it is as easy as this to tear down the cluster and release all of the resources: just delete the resource group. If you have deployed the cluster into an existing resource group, you’ll need to identify all resources that belong to the DC/OS cluster and manually delete them.

Next steps