Skip TestBuildContainerOOMKilled integration test

Romuald Atchadé - OOO until Jan 11th, 2026requested to merge
k8s-remove-failing-integration-test into main

What does this MR do?

Skip TestBuildContainerOOMKilled integration test

TestBuildContainerOOMKilled doesn't produce the same outputs depending on the Kubernetes cluster used:

  • On GKE 1.30.5-gke.1443001, the test passes without issue
  • On AWS EKS 1.31.0-20241024, the test fails
  • On the KIND cluster in the pipeline, the OOM seems to never happen
❯ go test -tags integration,kubernetes -timeout=60m -run "^TestBuildContainerOOMKilled*" -v gitlab.com/gitlab-org/gitlab-runner/executors/kubernetes
=== RUN   TestBuildContainerOOMKilled
=== RUN   TestBuildContainerOOMKilled/job_fails_because_build_container_is_OOMKilled
--- PASS: TestBuildContainerOOMKilled (16.26s)
    --- PASS: TestBuildContainerOOMKilled/job_fails_because_build_container_is_OOMKilled (13.76s)
PASS
ok      gitlab.com/gitlab-org/gitlab-runner/executors/kubernetes        (cached)
❯ go clean -cache
❯ go test -tags integration,kubernetes -timeout=60m -run "^TestBuildContainerOOMKilled*" -v gitlab.com/gitlab-org/gitlab-runner/executors/kubernetes
=== RUN   TestBuildContainerOOMKilled
=== RUN   TestBuildContainerOOMKilled/job_fails_because_build_container_is_OOMKilled
    kubernetes_integration_test.go:3590: 
                Error Trace:    /Users/ratchade/projects/main-runner/executors/kubernetes/kubernetes_integration_test.go:3590
                                                        /Users/ratchade/projects/main-runner/executors/kubernetes/kubernetes_integration_test.go:3609
                Error:          "\x1b[0KRunning with gitlab-runner development version (HEAD)\x1b[0;m\n\x1b[0K\x1b[36;1mPreparing the \"kubernetes\" executor\x1b[0;m\x1b[0;m\n\x1b[0;33mWARNING: Namespace is empty, therefore assuming 'default'.\x1b[0;m\n\x1b[0KUsing Kubernetes namespace: default\x1b[0;m\n\x1b[0KUsing Kubernetes executor with image alpine:3.14.2 ...\x1b[0;m\n\x1b[0KUsing attach strategy to execute scripts...\x1b[0;m\n\x1b[0K\x1b[36;1mPreparing environment\x1b[0;m\x1b[0;m\n\x1b[0KUsing FF_USE_POD_ACTIVE_DEADLINE_SECONDS, the Pod activeDeadlineSeconds will be set to the job timeout: 2h0m0s...\x1b[0;m\nWaiting for pod default/runner--project-0-concurrent-0-6mpath8c to be running, status is Pending\n\tContainersNotInitialized: \"containers with incomplete status: [init-permissions]\"\n\tContainersNotReady: \"containers with unready status: [build helper]\"\n\tContainersNotReady: \"containers with unready status: [build helper]\"\nRunning on runner--project-0-concurrent-0-6mpath8c via ee:7d:9c:1b:f5:65...\n\n\x1b[0K\x1b[36;1mGetting source from Git repository\x1b[0;m\x1b[0;m\n\x1b[32;1mFetching changes...\x1b[0;m\nInitialized empty Git repository in /builds/gitlab-org/ci-cd/gitlab-runner-pipeline-tests/gitlab-test/.git/\n\x1b[32;1mCreated fresh repository.\x1b[0;m\n\x1b[32;1mChecking out 69b18e5e as detached HEAD (ref is main)...\x1b[0;m\n\n\x1b[32;1mSkipping Git submodules setup\x1b[0;m\n\n\x1b[0K\x1b[36;1mExecuting \"step_script\" stage of the job script\x1b[0;m\x1b[0;m\n\x1b[32;1m$ echo \"Starting memory allocation to trigger OOM...\" # collapsed multi-line command\x1b[0;m\nStarting memory allocation to trigger OOM...\nKilled\n\n\x1b[31;1mERROR: Job failed: command terminated with exit code 137\n\x1b[0;m\n" does not contain "Error in container build: exit code: 137, reason: 'OOMKilled'"
                Test:           TestBuildContainerOOMKilled/job_fails_because_build_container_is_OOMKilled
--- FAIL: TestBuildContainerOOMKilled (8.53s)
    --- FAIL: TestBuildContainerOOMKilled/job_fails_because_build_container_is_OOMKilled (8.05s)
FAIL
FAIL    gitlab.com/gitlab-org/gitlab-runner/executors/kubernetes        9.229s
FAIL

Why was this MR needed?

Prevent integration_k8s from failing

What's the best way to test this MR?

integration_k8s job passes ✅

What are the relevant issue numbers?

None

Edited by Romuald Atchadé - OOO until Jan 11th, 2026

Merge request reports