EnterprisePrerequisite: DC/OS Enterprise cluster
You will need superuser access to an unutilized cluster of virtual machines (VMs). While more is usually always better for seeing more of what DC/OS Kubernetes works at scale, the following is the minimal setup for this current tutorial.
- 1 master node
- 5 private agent nodes
- 1 public agent node
- Each node with resources equivalent or greater to that of an
m4.xlargeAWS EC2 VM
Once you have your cluster infrastructure up and ready as specified by the above minimum prerequisites, you will next install both DC/OS CLIs:
- the DC/OS CLI
- the DC/OS Enterprise CLI Enterprise
before continuing on.
The Kubernetes HTTPS requirement:
The Mesosphere Kubernetes Engine (MKE) requires access over HTTPS in order to connect to the kubernetes-apiserver using
kubectl. So to finish setting up the cluster and before moving on to installing the MKE, let’s go ahead and make sure you are authenticated to the DC/OS CLI using HTTPS:
In your CLI, enter:
dcos config show core.dcos_url
If the returned URL does not start with
dcos config set core.dcos_url https://<dcos_master_public_IP_address>
Moreover, if the TLS certificate used by DC/OS is not trusted in your case, you can run the following command to disable TLS verification for the purposes of completing this tutorial:
dcos config set core.ssl_verify false
To verify that the cluster is connected, or if there was a change to
https, setup the cluster again, making sure to insert the actual URL of your cluster:
dcos cluster setup <cluster-url>
Accept any permissions requests for your certificate.
Next Step: Installing the Mesosphere Kubernetes Engine
Now that you have your DC/OS cluster all set up, you are ready to install the Mesosphere Kubernetes Engine.