Skip to content

LOCAL_TESTING.md

Latest commit

 

History

History
160 lines (116 loc) · 4.66 KB

LOCAL_TESTING.md

File metadata and controls

160 lines (116 loc) · 4.66 KB
Raw

Local Testing Guide

Run Tekton task tests locally for fast feedback before submitting PRs. The local environment mirrors CI exactly, ensuring tests that pass locally will pass in CI.

Quick Start

# One-time setup (5-10 minutes)
./scripts/local-test-env.sh setup
source .env.testing

# Run tests  
./scripts/run-local-tests.sh                    # Test changed files
./scripts/run-local-tests.sh --pr-mode          # Test PR changes
./scripts/run-local-tests.sh --parallel 2       # Parallel execution
./scripts/run-local-tests.sh tasks/managed/add-fbc-contribution  # Specific task

Prerequisites

Required tools: podman, kind, kubectl, yq, jq

Installation:

# macOS
brew install podman kind kubectl yq jq

# Fedora/RHEL  
sudo dnf install -y podman kubectl yq jq
curl -Lo ./kind https://kind.sigs.k8s.io/dl/v0.20.0/kind-linux-amd64
chmod +x ./kind && sudo mv ./kind /usr/local/bin/

# Ubuntu/Debian
sudo apt-get install -y podman jq
# (see setup script for kind, kubectl, yq installation)

The setup script will check for missing tools and guide installation.

Common Usage

# Test specific tasks
./scripts/run-local-tests.sh tasks/managed/add-fbc-contribution
./scripts/run-local-tests.sh tasks/managed/apply-mapping tasks/internal/update-fbc-catalog

# Test options
./scripts/run-local-tests.sh --dry-run                    # Preview what will be tested
./scripts/run-local-tests.sh --parallel 3                # Run 3 tests concurrently 
./scripts/run-local-tests.sh --remove-compute-resources   # For resource-constrained environments
./scripts/run-local-tests.sh --cleanup                    # Clean up after completion

# Results tracking
cat test-results/summary.txt                              # View test summary
ls test-results/                                          # Browse individual results

Environment Setup

# Setup creates a Kind cluster with Tekton and local registry
./scripts/local-test-env.sh setup

# Configure shell environment (run after setup and in each new shell)
source .env.testing

# Verify setup
kubectl cluster-info
tkn version

First-time setup takes 5-10 minutes. Subsequent testing sessions only require source .env.testing.

Environment Management

# Teardown environment 
./scripts/local-test-env.sh teardown              # Remove cluster (keep env files)
./scripts/local-test-env.sh teardown --all        # Remove everything including env files
./scripts/local-test-env.sh teardown --cluster-only # Remove only cluster

# Fresh restart
./scripts/local-test-env.sh reset                 # Teardown + setup with prompts
./scripts/local-test-env.sh reset --force         # Reset without prompts

Troubleshooting

Quick Fixes

Issue Solution
"kubectl context not set" source .env.testing
"Cannot access cluster" kind get clusters → re-run setup if missing
"Registry connection issues" Check port-forward: ps aux | grep 30001
Tests failing to start kubectl get pods -n tekton-pipelines

Debugging Test Failures

# View test logs
tkn pr list                           # List recent runs
tkn pr logs <pipelinerun-name>        # View specific test logs

# Check test results
cat test-results/summary.txt          # Overall results summary
cat test-results/<task-name>.log      # Individual task results

# Manual cleanup if needed
kubectl delete pipelinerun --all
kubectl delete taskrun --all

Advanced Usage

Workflow Modes

The test runner automatically detects and uses the appropriate workflow:

  • Trusted Artifacts - For tasks with ociStorage parameter (modern)
  • PVC - For legacy tasks using Persistent Volume Claims

Override with --trusted-artifacts or --pvc flags if needed.

Resource Management

# For resource-constrained environments
./scripts/run-local-tests.sh --remove-compute-resources

# Custom cluster name
export CLUSTER_NAME="my-test-cluster"
./scripts/local-test-setup.sh

Cleanup

# Clean up test resources only
kubectl delete pipelinerun --all
kubectl delete taskrun --all

# Remove entire cluster
kind delete cluster --name release-service-catalog

# Reset environment
rm -f .env.testing

Best Practices

  • Test locally first - Faster feedback than CI
  • Use --pr-mode before submitting PRs
  • Keep clusters running - Reuse between test sessions
  • Check results - Always review test-results/summary.txt
  • Parallel testing - Use --parallel N for independent tests

More Information