Skip to content

Add deploy_fma.sh and debug workflow for OCP E2E#357

Merged
diegocastanibm merged 9 commits intomainfrom
deploy-fma
Mar 19, 2026
Merged

Add deploy_fma.sh and debug workflow for OCP E2E#357
diegocastanibm merged 9 commits intomainfrom
deploy-fma

Conversation

@diegocastanibm
Copy link
Copy Markdown
Collaborator

@diegocastanibm diegocastanibm commented Mar 13, 2026

Description:

  • Extracts the inline deployment steps from ci-e2e-openshift.yaml into a reusable test/e2e/deploy_fma.sh script
  • Replaces ~386 lines of inline workflow YAML with a single call to the script
  • Adds --standalone flag for running outside CI (e.g., on a kind cluster)

This is a companion to #332. As discussed in that PR, we split the E2E workflow into two scripts:

  1. deploy_fma.sh (this PR) — deploys FMA infrastructure: CRDs, controllers (dual-pods + launcher-populator), test service
    account, test objects (ISC, LC, LPP, ReplicaSet), and verifies that pods are created, bound, and ready.
  2. ocp-test.sh (Standardize how to Run E2E tests in the OCP Test Infrastructure #332) — runs the actual E2E test scenarios (wake-up fast path, CEL policy validation, etc.) using the
    objects deployed by deploy_fma.sh.

Feature --standalone flag:

By default, deploy_fma.sh assumes the CI workflow has already created the namespace and GHCR pull secret. When --standalone is passed, the script handles namespace creation, cleanup, and pull secret setup, making it usable on any cluster with kubectl access.

Tested:

  • bash -n test/e2e/deploy_fma.sh syntax check passes
  • YAML validation passes on modified workflow
  • CI workflow (ci-e2e-openshift.yaml) runs successfully on PR trigger
  • Verify deploy_fma.sh outputs (INST, ISC, LC, LPP, RS) are correctly consumed by cleanup step

Signed-off-by: Diego-Castan <diego.castan@ibm.com>
Signed-off-by: Diego-Castan <diego.castan@ibm.com>
@diegocastanibm diegocastanibm marked this pull request as ready for review March 16, 2026 18:24
Signed-off-by: Diego-Castan <diego.castan@ibm.com>
… not referenced

  elsewhere)
  - Remove hardcoded Helm values that duplicate chart defaults (global.local=false,
  dualPodsController.sleeperLimit=2, launcherPopulator.enabled=true,
  dualPodsController.debugAcceleratorMemory=false)
  - Add HELM_EXTRA_ARGS env var to allow callers to pass arbitrary --set flags to helm upgrade
   --install
  - Update workflow to pass --set dualPodsController.debugAcceleratorMemory=false via
  HELM_EXTRA_ARGS
  - Show both controller images (dual-pods-controller + launcher-populator) in deploy step
  output

Signed-off-by: Diego-Castan <diego.castan@ibm.com>
@dumb0002
Copy link
Copy Markdown
Collaborator

@MikeSpreitzer, with the latest commits by @diegocastanibm, is this PR ready to be merged? Progress work for PR # 332 has dependency on this PR.

Signed-off-by: Diego-Castan <diego.castan@ibm.com>
Signed-off-by: Diego-Castan <diego.castan@ibm.com>
Copy link
Copy Markdown
Collaborator

@MikeSpreitzer MikeSpreitzer left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM (provided the Ci tests succeed)

@diegocastanibm diegocastanibm merged commit f8713af into main Mar 19, 2026
26 checks passed
@diegocastanibm diegocastanibm deleted the deploy-fma branch March 19, 2026 18:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants