Tasks

Tasks are the basic building block that you will use to build your larger pipelines. A Task is a collection of Steps that you define and arrange in a specific order of execution as part of your continuous integration flow. A Task executes as a Pod on your Kubernetes cluster. A Task is available within a specific namespace, while a ClusterTask is available across the entire cluster.

A Task declaration includes the following elements:

  • Parameters
  • Resources
  • Steps
  • Workspaces
  • Results


In a nutshell, tasks should perform a single operation in your CI/CD pipeline. Examples of tasks could include cloning a repository, compiling some code, or running a series of tests. When we build our own tasks, then we should also try to make them as reusable as possible.


We prepared our local environment to build, manage, and run Tekton CI/CD pipelines. It is now time to get started with some hands-on examples.


As a rule of thumb, your tasks will always start like this:

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: example-task-name
  labels:
     key: value
spec:


In the spec section, we can describe our actual task. This is where we can define the various elements we want to use to achieve the task's goal. Those elements could be parameters, workspaces, results, or steps.

Tasks are scoped to a single namespace in your cluster, but you might have ClusterTasks defined for the entire cluster by your administrators.

Once we trigger this task, everything will run inside a single Pod, which will be terminated once a step fails or once all of the steps are executed successfully.


Steps

Steps are the only required objects to create a task, and that makes sense. Steps describe the containers that will run as part of the task. This is where the actual operations to be performed on your inputs happen.

In the YAML file that describes the task, you define steps by adding an array describing the steps and the order in they should be performed in.

Each step must have, at a minimum, an image to use. It is also highly recommended to use a command value or a script field. This is because the container's entry point is overwritten with an executable that manages the step execution for Tekton.


A typical Step would also contain a name and would generally look like this:

spec:steps:- image: alpine:3.12command:- /bin/bash- -c- echo Text from a step


If you want to use an image from a private registry, you can add an ImagePullSecret Kubernetes object to the service account used by the task.


Now that you understand how tasks and Steps are defined, you will see how to use this knowledge to build your first working task.


First task Hello World task

this task might not be instrumental in your day-to-day life, it will demonstrate the basic concepts to build your first Tekton task:

First, start with a new YAML file called hello.yaml. In that file, write below content.

apiVersion: tekton.dev/v1beta1
kind: Task
metadata:
  name: hello
spec:
  steps:
   - name: 'print hello world'
     image: registry.access.redhat.com/ubi8/ubi-minimal
     command:
       - /bin/bash
       - -c
       - echo "Hello World"


Now that you have your first task defined, you can apply it to your cluster using the kubectl tool in your terminal:

$ kubectl apply –f ./hello.yaml


It is now time to use the tkn CLI tool to manage and execute this task. We can list the tasks that are available in the cluster by using task ls, as shown here:

learning-ocean:~ gaurav$ tkn task ls
NAME             DESCRIPTION   AGE
hello                          2 weeks ago


We can see the task that, just created here. Now that you know this task is available for you to use, you can start it by using the start command followed by the task's name:

learning-ocean:~ gaurav$ tkn task start hello 
TaskRun started: hello-run-88vs4
In order to track the TaskRun progress run:
tkn taskrun logs hello-run-88vs4 -f -n default

Notice that this created a TaskRun with a random name. The TaskRun, which is the actual execution of this task, is now running in your cluster. Because this task is running inside a Pod on the cluster, you can't see the logs directly in your console.


We can use the --showlog argument when we start the task to view the logs directly.

learning-ocean:~ gaurav$ tkn task start hello --showlog
TaskRun started: hello-run-x9q6k
Waiting for logs to be available...
[print hello world] Hello World
learning-ocean:~ gaurav$ 


This started a new task run with a new random name. This time, you will see the outputs of each step as they are executed.