Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

odo link flow broken on latest SBO 0.10. #5084

Closed
kadel opened this issue Sep 22, 2021 · 3 comments · Fixed by #5113
Closed

odo link flow broken on latest SBO 0.10. #5084

kadel opened this issue Sep 22, 2021 · 3 comments · Fixed by #5113
Assignees
Labels
estimated-size/M (10-20) Rough sizing for Epics. About 1 sprint of work for one person priority/Critical We need to stop doing everything and fix this ASAP. It is making the project unusable.
Projects

Comments

@kadel
Copy link
Member

kadel commented Sep 22, 2021

https://github.com/redhat-developer/service-binding-operator/releases/tag/v0.10.0

Latest SBO introduced changes that required explicitly create RoleBindings in order for SBO to be able to read information from CustomResources.

Ideally this RoleBinding should be automaticaly created when Operator that supports ServiceBinding gets installed.
But currently there are no Operator that would support SBO out of the box.

This means that when someone installs Operator that they want to use with SBO with, they need to create RoleBinding manually. This needs to be done with cluster admin privileges.

For example this is ClusterRole required for postresql from dev4devs operator.

apiVersion: rbac.authorization.k8s.io/v1
kind: ClusterRole
metadata:
  name: postgresql-binding
  labels:
    service.binding/controller: "true"
rules:
  - apiGroups:
      - postgresql.dev4devs.com
    resources:
      - databases
    verbs:
      - get
      - list

---

apiVersion: rbac.authorization.k8s.io/v1        
kind: ClusterRoleBinding
metadata:
  name: postgresql-binding-sbo
subjects:
- kind: ServiceAccount
  name: service-binding-operator
  namespace: openshift-operators
roleRef:
  kind: ClusterRole
  name: postgresql-binding
  apiGroup: rbac.authorization.k8s.io

We need to document this in odo docs. As without this users won't be able to use odo link with SBO.

/priority critical

@openshift-ci openshift-ci bot added the priority/Critical We need to stop doing everything and fix this ASAP. It is making the project unusable. label Sep 22, 2021
@kadel
Copy link
Member Author

kadel commented Sep 22, 2021

Something that we could consider is to drop SBO use form odo, and just use the client side library (linking without operator), even if the SBO is installed on the cluster.

Currently there is no real benefit for using the Service Binding Operator. There are no operators that are using SBO, so user needs to define binding on their own anyway.

Not supporting the operator approach for now would also have a good inpact on odo link stability, as SBO api is unstable and it is breaking with every release.
This would be just temporarly, once SBO api is stable and usable, and there are operators that can be used out of the box, we would enable option for using the Operator again.

@kadel kadel added this to For Consideration in Sprint 208 via automation Sep 23, 2021
@dharmit dharmit moved this from For Consideration to To Do in Sprint 208 Sep 27, 2021
@scottkurz
Copy link
Contributor

One more note: IIUC, the SBO is not correctly watching the managed CR for changes:
redhat-developer/service-binding-operator#970
taking away yet another benefit for using SBO, until this is fixed.

@feloy
Copy link
Contributor

feloy commented Sep 29, 2021

  • Disable the test that SBO is running on the cluster, always use the library
  • Keep the code for SBO use, we will use it later
  • Remove SBO from clusters

@dharmit dharmit added the estimated-size/M (10-20) Rough sizing for Epics. About 1 sprint of work for one person label Sep 29, 2021
@feloy feloy moved this from To Do to In Progress in Sprint 208 Oct 4, 2021
@feloy feloy moved this from In Progress to For Review in Sprint 208 Oct 5, 2021
Sprint 208 automation moved this from For Review to Done Oct 5, 2021
@kadel kadel mentioned this issue Nov 10, 2021
4 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
estimated-size/M (10-20) Rough sizing for Epics. About 1 sprint of work for one person priority/Critical We need to stop doing everything and fix this ASAP. It is making the project unusable.
Projects
No open projects
Development

Successfully merging a pull request may close this issue.

4 participants