-
Initiative
-
Resolution: Unresolved
-
Critical
-
None
-
None
-
Proactive Architecture
-
False
-
-
False
-
15% To Do, 15% In Progress, 70% Done
-
4
-
0
Feature Overview
Hosted Control Planes (aka HyperShift) is an OpenShift form factor that optimizes for cost and time-to-provision as well as portability cross-cloud with a strong separation of concerns between management and workloads(for more information about HyperShift, check [1]). While HyperShift is OpenShift, it is still a different technical architecture, as a result, it becomes achieves functional parity with existing standalone OpenShift architecture and part of this effort is to ensure that HyperShift deployed OpenShift is compatible with existing layered operators.Â
Main StoryÂ
When deploying and life cycling OpenShift clusters with Hosted Control Planes (aka HyperShift), I want to deploy any supported OpenShift layered operator to fully utilize OpenShift value-add.
RequirmentsÂ
- Important layered operators are supported first (operators included in OKE/OCP/OPP)
- All other layered operators should be qualified on HyperShift by their respective teams
Additional Information
The goal of this Jira is to make sure all layered operators are tested/validated on Hosted OpenShift clusters (HCP, created by the HyperShift operator), in the same way they are with standalone OpenShift clusters.Â
IF testing failed, here is a few key points to keep in mind. Hosted OpenShift clusters don't have master nodes, and they don't expose Machine Config to cluster instance admins (see personas). A quick list of differences can be found here: https://docs.openshift.com/container-platform/4.17/hosted_control_planes/index.html#hcp-ocp-differences_hcp-overviewÂ
To get your operator to work with HCP, here are some quick things to ensure:
- Your operator does not hard-code to master nodes (and does not depend on them).
- Your operator logic does not modify MachineConfigs directly.Â
Â
Also, feel free to reach out on #project-hypershift for questions.Â
OutcomeÂ
The compatibility of all layered operators is documented in [2]. Also results of operator using RBAC to read/write to MachineConfig/MachineConfigPool thanks to the operator framework team.
ReferencesÂ
- depends on
-
OCPSTRAT-1427 Ability to run CRO on infra/worker node ( this will also enable CRO to run in HCP)
-
- Closed
-
- is related to
-
CMP-1234 HyperShift Support for Compliance Operator
-
- Closed
-
-
RFE-5592 External DNS operator doesn't work on HCP clusters
-
- Accepted
-
-
OCPBUGS-37059 [ExternalDNS] Unable to deploy ExternalDNS on ROSA HCP cluster
-
- Closed
-
-
PODAUTO-65 Failed to install Vertical Pod Autoscaler Operator on Hypershift guest cluster
-
- Closed
-
-
OCPBUGS-23172 Add a release note which reads RODO could not be installed on Hypershift
-
- Closed
-
- relates to
-
OCPBUGS-3410 Failed to install File Integrity Operator on Hypershift guest cluster
-
- New
-
-
OCPSTRAT-1509 Tokenized Auth Enablement for OLM-managed Operators on Azure
-
- New
-
-
OCPSTRAT-6 Tokenized Auth Enablement for OLM-managed Operators on AWS
-
- In Progress
-
-
OCPSTRAT-127 Continued STS enablement for selected OLM-managed operators
-
- In Progress
-
-
OCPSTRAT-103 Ensuring the Control Plane is Fully Decoupled for Hosted Control Planes
-
- Closed
-
- links to