Running custom kubectl commands

How to use kubectl in your Codefresh pipelines

As explained in the deployment options page, Codefresh has several built-in facilities for deploying to Kubernetes clusters.

If you wish you can still run your own custom kubectl commands in a freestyle step for maximum flexibility on cluster deployments. Kubectl is the command line interface for managing kubernetes clusters.

Codefresh helps you even in this scenario by automatically setting up your config context with your connected clusters.

The config context is automatically placed for you at the path of the variable $CF_KUBECONFIG_PATH. In the current Codefresh implementation this expands to /codefresh/volume/sensitive/.kube/config, inside the shared step volume.

Notice that when you use custom kubectl commands, it is your responsibility to template your manifests using any of the available options. If you wish to employ Codefresh for templating it is better to use the dedicated cf-deploy-kubernetes step which provides simple templating capabilities.

Using the Codefresh kubectl image

Codefresh already offers a public docker image with kubectl at https://hub.docker.com/r/codefresh/kubectl/tags. You can choose a specific version of kubectl with the appropriate tag or just select latest of the most up-to-date version.

YAML

version: '1.0'
steps:
  MyCustomKubectlCommands:
    title: Running Kubectl
    image: codefresh/kubectl:1.13.3
    commands: 
      - echo $CF_KUBECONFIG_PATH
      - kubectl help

If you run the pipeline you will see the help options for kubectl.

Getting a config context

The important thing to know when running custom kubectl commands is that Codefresh is automatically setting up your kubeconfig files for you with the cluster information present in integrations

If you run this pipeline, you will see the names of all your connected clusters

YAML

version: '1.0'
steps:
  MyCustomKubectlCommands:
    title: Running Kubectl
    image: codefresh/kubectl
    commands: 
      - kubectl config get-contexts

With two sample clusters the output of this pipeline is the following:

Running freestyle step: Running Kubectl
Pulling image codefresh/kubectl:latest
Status: Image is up to date for codefresh/kubectl:latest
NAME                              CLUSTER                           AUTHINFO                          NAMESPACE
gke-kostisdemo-codefresh-kostis   gke-kostisdemo-codefresh-kostis   gke-kostisdemo-codefresh-kostis   default
[email protected]       [email protected]stKubernetes       [email protected]       default

You can modify the current config context and run any kubectl command you want applied to that context. The next pipeline will print all the nodes of the first cluster:

YAML

version: '1.0'
steps:
  MyCustomKubectlCommands:
    title: Running Kubectl
    image: codefresh/kubectl
    commands: 
      - kubectl config get-contexts
      - kubectl config use-context "gke-kostisdemo-codefresh-kostis"
      - kubectl get nodes

Example of parallel deployment with kubectl

Let’s see a full example. In this pipeline we are going to create two docker images and deploy them into two separate clusters, using custom kubectl commands. We are going also to use the parallel capability of Codefresh pipelines.

Here is the pipeline:

and here is the full codefresh.yml.

YAML

version: '1.0'

stages:
- build
- deploy

steps:
  BuildingApps:
    type: parallel
    stage: 'build'
    steps:
      BuildingApp1:
        title: Building App 1
        type: build
        stage: build
        image_name: nestjs-app
        working_directory: ./my-nestjs-project/
        dockerfile: Dockerfile
      BuildingApp2:
        title: Building App 2
        type: build
        stage: build
        image_name: rails
        working_directory: ./my-rails-project/
        dockerfile: Dockerfile
  DeployingApps:
    type: parallel
    stage: 'deploy'
    steps:
      DeployApp1:
        title: Deploying App 1
        stage: deploy
        image: codefresh/kubectl
        working_directory: ./my-nestjs-project/
        commands: 
          - kubectl config get-contexts
          - kubectl config use-context "gke-kostisdemo-codefresh-kostis"
          - kubectl apply -f service.yml deployment.yml
      DeployApp2:
        title: Deploying App 2
        stage: deploy
        image: codefresh/kubectl
        working_directory: ./my-rails-project/
        commands: 
          - kubectl config get-contexts
          - kubectl config use-context "[email protected]"  
          - kubectl apply -f service.yml deployment.yml configmap.yml

In this example above we select the one of the clusters on each deployment step, and then apply several Kubernetes manifests that constitute an application.

Running custom kubectl commands

Using the Codefresh kubectl image

Getting a config context

Example of parallel deployment with kubectl

What to read next