make GC mutation check ignore selfLink

[deads2k]

fixes #15044

Got a reproducer here https://bugzilla.redhat.com/attachment.cgi?id=1295674 and for some reason the selfLink was causing the GC mutation check to fail.

@mfojtik do you have someone itching to learn about GC?

[test]

[openshift-bot]

Evaluated for origin test up to afbed8d

[openshift-bot]

continuous-integration/openshift-jenkins/test SUCCESS (https://ci.openshift.redhat.com/jenkins/job/test_pull_request_origin/3030/) (Base Commit: 18c5730) (PR Branch Commit: afbed8d)

[ironcladlou]

LGTM for the specific problem case, but I still wonder if this is comprehensive. Consider a scenario where the request obj omits certain immutable metadata fields like creationTimestamp- wouldn't the old object from storage have creationTimestamp populated, resulting in another false positive diff?

[deads2k]

Consider a scenario where the request obj omits certain immutable metadata fields like creationTimestamp- wouldn't the old object from storage have creationTimestamp populated, resulting in another false positive diff?

That wouldn't cause a problem on patches. This fails for patches which aren't trying to modify anything else.

[merge][severity:blocker]

[liggitt]

flaked on

Expected
    <int>: 0
to be >
    <int>: 0
/go/src/github.com/openshift/origin/vendor/k8s.io/kubernetes/test/e2e/daemon_set.go:445

[merge]

[deads2k]

re[merge]

[openshift-bot]

Evaluated for origin merge up to afbed8d

[openshift-bot]

continuous-integration/openshift-jenkins/merge SUCCESS (https://ci.openshift.redhat.com/jenkins/job/merge_pull_request_origin/1273/) (Base Commit: 14f47f9) (PR Branch Commit: afbed8d) (Extended Tests: blocker) (Image: devenv-rhel7_6438)