Bug 1891993

Is there a workaround for the OCP 4.6 operators to show up?

Hello,

I have received following report from our partner NEC which states similar issue, would you please find it below ?

Feedback from NEC:
~~~
We have found the cause of the problem. If RH has already created Bugzilla, Would you please share the following with engineers ?

After the OCP4.6 upgrade, we still have old pods on openshift-marketplace.

  $ oc get pods -n openshift-marketplace
  NAME                                    READY   STATUS    RESTARTS   AGE
  certified-operators-69c7584d8c-z9vjr    1/1     Running   0          3d22h <-- For OCP4.5
  certified-operators-jdbt2               1/1     Running   0          2d7h
  community-operators-54577646b6-qn9l5    1/1     Running   0          3d22h <-- For OCP4.5
  community-operators-bttkx               1/1     Running   0          14h
  marketplace-operator-67d858bf7d-24r29   1/1     Running   0          3d22h
  redhat-marketplace-4wsts                1/1     Running   0          2d7h <-- For OCP4.5
  redhat-marketplace-55c6c798f9-2twk6     1/1     Running   0          3d22h
  redhat-operators-6d98f9c45c-xqht7       1/1     Running   0          3d22h <-- For OCP4.5
  redhat-operators-8nscg                  1/1     Running   0          3d22h

Apparently, since OCP4.6, these pods are controlled directly by the Operator, but the The old deployment information used in OCOP4.5 and earlier is still there, and this makes it possible to use the seems to be occurring

  $ oc get deployment -n openshift-marketplace
  NAME                   READY   UP-TO-DATE   AVAILABLE   AGE
  certified-operators    1/1     1            1           117d
  community-operators    1/1     1            1           117d
  marketplace-operator   1/1     1            1           490d <== OCP4.6 fresh env has only this
  redhat-marketplace     1/1     1            1           117d
  redhat-operators       1/1     1            1           117d

After the following actions, the 4.6 operators are now available: 1.

  1. remove all openshift-marketplace deployments

    $ oc delete deployment -n openshift-marketplace certified-operators community-operators redhat-marketplace redhat-operators marketplace-operator

  2. wait until the marketplace-operator is created

  3. disable all CatalogSource and then re-enable them.

    $ oc patch OperatorHub cluster --type json -p '[{"op": "add", "path": "/spec/disableAllDefaultSources", "value": true}]'
    operatorhub.config.openshift.io/cluster patched
    $ oc patch OperatorHub cluster --type json  -p '[{"op": "add", "path": "/spec/disableAllDefaultSources", "value": false}]'
    operatorhub.config.openshift.io/cluster patched

~~~

Dear Red Hat,

Please keep in mind that it's not only a problem which consumes resources.
User cannot install any Operators of 4.6 version due to this issue, since '4.6' channel is hidden by the old resources.

 $ oc get packagemanifest cluster-logging -n openshift-marketplace -o jsonpath='{.status.channels[*].name}'
 4.2 4.2-s390x 4.3 4.4 4.5 preview

We don't know why Red Hat couldn't find this issue before releasing 4.6 as GA, since it would be able to find easily with a basic upgrade test.
Anyway, we hope that Red Hat fix this issue early.

Best Regards,
Masaki Hatada

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (Moderate: OpenShift Container Platform 4.7.0 security, bug fix, and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHSA-2020:5633