-
Bug
-
Resolution: Done
-
Normal
-
None
-
4.20
-
Quality / Stability / Reliability
-
False
-
-
None
-
Moderate
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
-
None
Description of problem:
The `A successfully built MachineOSConfig can be re-build` OnClusterBuild test is flaky and has a 7-day pass rate below 95%.
Version-Release number of selected component (if applicable):
Seen in 4.20, possible in other versions.
How reproducible:
Failures happen intermittently, I have not been able to reproduce locally.
Steps to Reproduce:
See the following test runs to see examples of the test failures:
- e2e-aws-mco-disruptive-techpreview
- e2e-metal-ipi-ovn-ipv4-mco-disruptive-techpreview
- e2e-vsphere-mco-disruptive-techpreview
Actual results:
The `e2e-aws-mco-disruptive-techpreview` test failed with the following error:
fail [github.com/openshift/machine-config-operator/test/extended/mco_ocb.go:355]: mc operation is not completed on mcp worker: mcp worker has degraded 1 machines Unexpected error: <*errors.errorString | 0xc000685e50>: mcp worker has degraded 1 machines { s: "mcp worker has degraded 1 machines", } occurred
The remaining failures had the following failure:
fail [github.com/openshift/machine-config-operator/test/extended/mco_ocb.go:423]: OCL was disabled in ci-op-t5sllq3v-40d0b-qcnrc-worker-0-5xlll but the OCL image is still used in %!s(MISSING) Expected <string>: image-registry.openshift-image-registry.svc:5000/openshift-machine-config-operator/ocb-worker-image@sha256:48bf43eb3c20860a63e5a5bf21429ca0c51267d98f21db30f73eac2cb4f50dcf not to equal <string>: image-registry.openshift-image-registry.svc:5000/openshift-machine-config-operator/ocb-worker-image@sha256:48bf43eb3c20860a63e5a5bf21429ca0c51267d98f21db30f73eac2cb4f50dcf
Expected results:
The test should consistently pass.
Additional info:
N/A
