Slice 3: Create airgap test pipeline (Jenkinsfile.airgap-rke2-tests)

[floatingman]

Parent PRD

#585

What to build

Create a Declarative Pipeline Jenkinsfile.airgap-rke2-tests that replaces Jenkinsfile.airgap.go-tests (405 lines) — the most comprehensive airgap pipeline that combines setup, test execution, and teardown into a single pipeline.

This pipeline consumes the shared functions added in #589. It uses make.runTarget() for cluster/registry/rancher operations (same as Slice 2) and shares the same infrastructure setup stages.

Full lifecycle flow:

1. Standard dual-repo checkout via airgap.standardCheckout
2. Build Dockerfile.airgap-go-tests image (kept separate from Dockerfile.infra to preserve build time)
3. Setup infrastructure (same stages as Slice 2 setup flow, using shared functions):
   • Tofu init → workspace create → apply
   • Generate inventory from tofu outputs
   • Configure Ansible variables via airgap.configureAnsible
   • Copy SSH keys to nodes via ansible.runPlaybook
   • Deploy RKE2 cluster via make.runTarget(target: 'cluster')
   • Configure private registry via make.runTarget(target: 'registry') (conditional)
   • Deploy Rancher via make.runTarget(target: 'rancher') (unconditional in this pipeline)
4. Generate cattle-config.yaml — retrieve tofu outputs, run admin token generation playbook
5. Run Go test suite via gotestsum with standardized flags and output format
6. Optionally report results to Qase (controlled by QASE_TEST_RUN_ID parameter)
7. Teardown infrastructure if DESTROY_AFTER_TESTS=true (via post { always })
8. Teardown infrastructure on failure if DESTROY_ON_FAILURE=true (via post { failure })

Key difference from Slice 2: This pipeline always deploys Rancher (no DEPLOY_RANCHER toggle), includes test execution stages, and has the DESTROY_AFTER_TESTS option. The teardown logic uses the same airgap.teardownInfrastructure function as the destroy pipeline.

Parameters (harmonized):

• TEST_PACKAGE — Go test package to run
• CONFIG — Test configuration file
• DESTROY_ON_FAILURE (boolean, default: true)
• DESTROY_AFTER_TESTS (boolean, default: true)
• QASE_TEST_RUN_ID (string, optional) — enables Qase reporting when set
• Standard branch parameters, version parameters, TERRAFORM_CONFIG, ANSIBLE_VARIABLES

Acceptance criteria

• Jenkinsfile.airgap-rke2-tests created in Declarative Pipeline syntax
• Infrastructure setup uses same shared functions as Slice 2
• RKE2 deployment uses make.runTarget(target: 'cluster')
• Rancher deployment is unconditional (always runs via make.runTarget(target: 'rancher'))
• cattle-config generation uses shared token extraction pattern
• gotestsum invocation uses standardized flags and output format
• JUnit XML and JSON test results are archived
• Qase reporting runs as optional stage when QASE_TEST_RUN_ID is set
• DESTROY_ON_FAILURE triggers teardown on any stage failure via post { failure }
• DESTROY_AFTER_TESTS triggers teardown after successful run via post { always }
• Teardown uses same airgap.teardownInfrastructure as Slice 2 destroy
• All shared functions from Slice 1: Add shared pipeline functions to qa-jenkins-library #589 are consumed
• Original Jenkinsfile.airgap.go-tests remains untouched (parallel coexistence)
• Pipeline passes at least 2 successful live test execution cycles

Blocked by

• Blocked by Slice 1: Add shared pipeline functions to qa-jenkins-library #589 (shared pipeline functions must be available) Slice 1: Add shared pipeline functions to qa-jenkins-library #589 is complete.
• Can proceed once Slice 2 (Slice 2: Create unified airgap infra pipeline (Jenkinsfile.airgap-rke2-infra) #590) setup patterns are validated.

User stories addressed

• User story 5 (tofu lifecycle via shared function)
• User story 7 (checkout via shared function)
• User story 9 (go-tests pipeline shares same checkout and infra functions)
• User story 10 (standardized gotestsum invocation)
• User story 11 (Qase reporting as optional stage)
• User story 12 (teardown logic shared with destroy pipeline)
• User story 13 (cattle-config generation as shared function)
• User story 21 (Declarative Pipeline syntax)
• User story 22 (parallel coexistence)
• User story 23 (simplified naming)
• User story 24 (harmonized parameters)
• User story 22-24 (Makefile integration for cluster/registry/rancher)