-
Bug
-
Resolution: Done
-
Major
-
None
Description of problem:
If a problem prevents the hypershift-install-job from completing, it does not retry sufficiently enough to overcome the problem.
Version-Release number of selected component (if applicable):
advanced-cluster-management.v2.7.0-rc03
How reproducible:
The hypershift-install-job may fail, e.g:
oc logs hypershift-install-job-nz66v-wgkvb already exists: PriorityClass /hypershift-control-plane already exists: PriorityClass /hypershift-etcd already exists: PriorityClass /hypershift-api-critical already exists: PriorityClass /hypershift-operator applied Namespace /hypershift applied ServiceAccount hypershift/operator applied ClusterRole /hypershift-operator applied ClusterRoleBinding /hypershift-operator applied Role hypershift/hypershift-operator applied RoleBinding hypershift/hypershift-operator applied ServiceAccount hypershift/external-dns applied ClusterRole /external-dns applied ClusterRoleBinding /external-dns applied Deployment hypershift/external-dns applied Deployment hypershift/operator applied Service hypershift/operator applied Role hypershift/prometheus applied RoleBinding hypershift/prometheus Error: no matches for kind "ServiceMonitor" in version "monitoring.rhobs/v1" no matches for kind "ServiceMonitor" in version "monitoring.rhobs/v1"
In that case the job is retried 3 times. If neither of those additional attempts succeeds, the installation does not proceed further.
Actual results:
Even once the underlying problem gets fixed (in this case installation of RHOBS CRDs is complete), Hypershift operator installation never finishes.
Expected results:
The installation job should keep retrying until installation is successful.
Additional info: