Guided Tutorial: CLI

This tutorial is a continuation of the Guided Tutorial and you should start there.


You will be creating a single Pod in this tutorial which is not sufficient for a production service

Adding a Pod to your Project

Before using the CLI it would be wise to read our Getting Started on the CLI doc.

Once the oc client has been installed and is logged into the cluster you need to switch to your Project. Switching to a Project allows the oc client to assume that the commands it is running should be executed inside of the Project that you switch to. You could alternatively not switch to a project and append the -n flag to each command you run followed by the name of the project you wish to run your command in. That being said, switch to your project:

oc project <PROJECT_NAME>

Where <PROJECT_NAME> above is the name of your Project which will be the name of your RATS allocation. If you need to double check your Project name you can run:

oc get projects

to get a list of projects that you have access to.

Now, to get a list of pods that exist in the project run:

oc get pods

This should not return any Pods because there will not be any pods in the project yet.

To remedy this problem we will create a pod. Below is a basic definition of a pod in YAML. Copy this and save it to a file named pod.yaml.

apiVersion: v1
kind: Pod
  # Pod name
  name: test-pod
    # Container name
    - name: test-container
      # Using the base image
      image: "image-registry.openshift-image-registry.svc:5000/openshift/ccs-rhel7-base-amd64"
      # Starting a shell
      command: ["/bin/sh","-c"]
      # Echoing a Hello World followed by an infinitely waiting cat
      args: ["echo 'Hello World!'; cat"]
      # Need a tty if we are to SSH. Need stdin for tty
      tty: true
      stdin: true

With the above YAML saved in a pod.yaml file we can now create the pod with the following command:

oc create -f pod.yaml

Now if we run the command:

oc get pods

We should see our pod along with some status information about the pod.

To get useful metrics about the pod we can run:

oc describe pod <POD_NAME>

where <POD_NAME> will be the name of the pod. In our case test-pod.

To get logs from the pod we can run the command:

oc logs -f <POD_NAME>

NOTE the -f flag will follow the logs. You can run the logs command without the -f flag to get a snapshot of the logs.

oc get pod <POD_NAME> -o yaml

Will allow you to view the YAML representation that exists in Openshift that defines your pod. You may notice that the YAML contains many more key/value pairs than the YAML that we have in our pod.yaml file. This is correct and is because extra YAML is added during the pod creation process.

Finally, to get a remote shell in the pod we run the oc rsh <POD_NAME> command. This will default to using /bin/sh in the pod. If a different shell is required, we can provide the optional --shell=/path/to/shell flag. For example, if we wanted to open a bash shell in the pod we would run the following command:

oc rsh --shell='/bin/bash' <POD_NAME>

If you have multiple containers in your pod, the oc rsh <POD_NAME> command will default to the first container. If you would like to start a remote shell in one of the other containers, you can use the optional -c <CONTAINER_NAME> flag.