Closed mudler closed 4 years ago
Please verify if #1605 is fixing this issue as well.
Given that the cluster role definition works when Eirini is deployed only once, it feels like the role definition is correct.
But since the role binding is using a static global name, the second install may be breaking the first one by modifying its role binding. #1605 prepends cluster role and binding names with the release namespace name, so there should be no more name collision.
@divyaaswath #1605 has been merged to master
.
I hope it fixes https://github.com/cloudfoundry-incubator/quarks-operator/issues/1227 for you; please re-open this bug if it doesn't!
Describe the bug By deploying Eirini twice in the same cluster, the following is observed in the eirini logs:
and the cloud-controller-clock container in the scheduler pod crashes because of this.
To Reproduce Deploy two KubeCF with Eirini and 1 operator
Environment
Additional context https://github.com/cloudfoundry-incubator/quarks-operator/issues/1227