-
Bug
-
Resolution: Done-Errata
-
Major
-
None
-
4.16
-
+
-
Moderate
-
None
-
NE Sprint 256, NE Sprint 257, NE Sprint 258
-
3
-
Rejected
-
False
-
Description of problem:
The ExternalDNS Operator encounters problems during the deployment of operands on HCP clusters (ROSA and vanilla HyperShift). - The operand deployment's node placement is restricted to control plane nodes (https://github.com/openshift/external-dns-operator/blob/48345f6ac0a7ed3d76b6d7532f8b88d9222990d2/pkg/operator/controller/externaldns/deployment.go#L203-L214) while the HCP cluster only have nodes with the worker profile. - The operand deployment does not have the bound service account token mounted inside the container. This prevents the use of STS-enabled credential secrets. A similar bug was previously fixed in ALBO: https://github.com/openshift/aws-load-balancer-operator/pull/51.
Version-Release number of selected component (if applicable):
All GA versions (1.x)
How reproducible:
Always
Steps to Reproduce:
1. Create HyperShift HCP or ROSA HCP cluster 2. Install ExternlDNS Operator from the OperatorHub 3. Create a given sample of ExternalDNS custom resource: apiVersion: externaldns.olm.openshift.io/v1beta1 kind: ExternalDNS metadata: name: sample-aws spec: provider: type: AWS aws: credentials: name: aws-sts-creds source: type: OpenShiftRoute openshiftRouteOptions: routerName: default zones: - Z04015592QJX3EK1YYYY
Actual results:
ExternalDNS deployment stuck in Pending state.
Expected results:
ExternalDNS deployment is Running and ready.
Additional info:
- incorporates
-
RFE-5592 External DNS operator doesn't work on HCP clusters
- Accepted
- is related to
-
RFE-5408 ExternalDNS Operator: expose pod nodeSelector in the CRD
- Backlog
-
RFE-5592 External DNS operator doesn't work on HCP clusters
- Accepted
- relates to
-
OCPSTRAT-605 Ensure compatibility of layered operators for HCP (HyperShift)
- In Progress
- links to
-
RHEA-2024:139597 ExternalDNS Operator 1.3 operator/operand containers
- mentioned on
(2 links to, 9 mentioned on)