-
Bug
-
Resolution: Obsolete
-
Major
-
None
-
None
Description of problem:
1-010_validate-ootb-manage-other-namespace Kuttl test sometimes fails with flaky behaviors Openshift CI.
In the last few weeks, the test has failed around 26 times across 10 PRs with re-runs. It has been seen on all three OCP versions
Flake board: https://github.com/anandrkskd/gitops-operator-flake-board/blob/main/output/flake-stats.md
Prerequisites (if any, like setup, operators/versions):{}
Steps to Reproduce
Presubmit jobs run on each PR
Actual results:
Most of the times the failure occurred when resources that were expected to be deleted still existed in the cluster
case.go:361: failed in step 5-check-rbac case.go:363: resource matched of kind: rbac.authorization.k8s.io/v1, Kind=Role case.go:363: resource matched of kind: rbac.authorization.k8s.io/v1, Kind=Role case.go:363: resource matched of kind: rbac.authorization.k8s.io/v1, Kind=RoleBinding
Other failures seem to be because of the previous test failures (1-006_validate_machine_config and 1-018_validate_disable_default_instance) :
test step failed 1-install resource ArgoCD:openshift-gitops/openshift-gitops: .status.phase: key is missing from map
and
case.go:361: failed in step 1-install case.go:363: argocds.argoproj.io "openshift-gitops" not found
Expected results:
The test should succeed
Reproducibility (Always/Intermittent/Only Once):
Intermittent
Dependency
For the last failure (test step failed 1-install) it should probably be fixed with
https://issues.redhat.com/browse/GITOPS-2844
https://issues.redhat.com/browse/GITOPS-2829
Acceptance criteria:
- Above mentioned test should succeed in Openshift CI
Definition of Done:
- Acceptance criterias are met
- Openshift CI does not encounter any flaky tests
Build Details:
Additional info (Such as Logs, Screenshots, etc):
*
- clones
-
GITOPS-2857 Test failure | 1-055_validate_notification_controller
- Closed
- is cloned by
-
GITOPS-2998 Test failure | 1-035_validate_argocd_secret_repopulate
- Closed
- mentioned on