Sub-pipeline example: Pipeline with multiple tasks

Source code for this example.

• In the Radix application repository create a folder tekton. This folder need to be in the configuration branch and in the same folder, where radixconfig.yaml file is located (by default it is a root of the repository).

• The sub-pipeline in this example runs multiple tasks.

• Create files with tasks. File names of tasks can be arbitrary (no any connection with a task actual name, used in a sub-pipeline), file extensions should be yaml.

  • Create a file hello-task.yaml for the task hello. This task runs in the container with Alpine Linux and execute one command echo "Hello".

    apiVersion: tekton.dev/v1
    kind: Task
    metadata:
      name: hello
    spec:
      steps:
        - name: say-hello
          image: alpine
          command:
            - echo
          args:
            - "Hello"

  • Create a file say-letters-task.yaml for the task letters. This task runs in the container with a Python, within Alpine Linux and execute a script, put directly in the task description. The script prints few letters with 1 second interval.

    apiVersion: tekton.dev/v1
    kind: Task
    metadata:
      name: letters
    spec:
      steps:
        - name: say-letters
          image: alpine
          script: |
            #!/bin/sh
            for i in "a" "b" "c" "d" "e"
              do echo $i
              sleep 1
            done
            echo "Done"

  • Create a file say-numbers-task.yaml for the task numbers. This task is similar to the task letters, but its script prints few numbers with 2 second interval. This task runs longer then letters task to see running them in parallel.

    apiVersion: tekton.dev/v1
    kind: Task
    metadata:
      name: numbers
    spec:
      steps:
        - name: say-numbers
          image: alpine
          script: |
            #!/bin/sh
            for i in "one" "two" "three" "four" "five" "six" "seven" "eight" "nine" "ten"
              do echo $i
              sleep 2
            done
            echo "Done"

  • Create a file goodbye-task.yaml for the task goodbye. This task runs in the container with a Python on Alpine Linux. The task executes a inline Python command print('Goodbye').

    apiVersion: tekton.dev/v1
    kind: Task
    metadata:
      name: goodbye
    spec:
      steps:
        - name: say-goodbye
          image: python:alpine
          command:
            - python
          args:
            - "-c"
            - "print('Goodbye')"

• Create a file pipeline.yaml. Add tasks in the tasks list: give them names (it can be any name, unique within this sub-pipeline), in the property taskRef ("reference to a task") put the value from the property metadata.name of the tasks, created above:

apiVersion: tekton.dev/v1
kind: Pipeline
metadata:
  name: pipeline-with-multiple-steps
spec:
  tasks:
    - name: show-hello
      taskRef:
        name: hello

    - name: show-goodbye
      runAfter:
        - show-numbers
        - show-letters
      taskRef:
        name: goodbye

    - name: show-letters
      runAfter:
        - show-hello
      taskRef:
        name: letters
    - name: show-numbers
      runAfter:
        - show-hello
      taskRef:
        name: numbers

• File structure can be like this:

/
├── tekton/
│   ├── pipeline.yaml
│   ├── hello-task.yaml
│   ├── say-letters-task.yaml
│   ├── say-numbers-task.yaml
│   └── goodbye-task.yaml
└── radixconfig.yaml

This sub-pipeline runs first the task show-hello (which reference to the task hello described in the file hello-task.yaml), then two tasks in parallel show-numbers and show-letters (referenced to the tasks numbers and letter correspondingly) - these tasks wait for completion of the task show-hello as both have set the optional property runAfter

runAfter:
  - show-hello

The last task to be running is show-goodbye (which reference to the task goodbye described in the file goodbye-task.yaml). This tak is started after completion of both tasks show-numbers and show-letters, as it has set the optional property runAfter, referenced to both these tasks

runAfter:
  - show-numbers
  - show-letters

It is not important in which order to put the tasks - tasks can run in parallel or in a sequences, defined by fields runAfter, conditions, from.

• Commit changes in the repository. Look at the details of a started Radix pipeline job (if the Radix app is connected to the GitHub WebHook, otherwise - start a job manually).
• Navigate to the Radix pipeline step "Run pipeline", when it is running or completed: the pipelines overview page shows a table with a list of sub-pipelines - in this example it is one sub-pipeline "pipeline-example", running for an environment "dev", and the sub-pipeline status.
• Navigate to the sub-pipeline (click on its name in the table row)
• The pipeline overview page shows a table with a list of this sub-pipeline's tasks - in this example it is a list of tasks "show-hello", "show-letters", "show-numbers", "show-goodbye", and the task statuses.
• Navigate to the task (click on its name in the table row)
• The sub-pipeline task overview page shows a table with a task's step, with a name of a step, step status and the step log.