[Doc] for Submariner 0.9+ operator image should be accessible before broker deploy

[manosnoam]

Since the merge of submariner-io/submariner-operator#1085
If using SubCtl with the override-images option, it is required to set access to those images before the subctl deploy broker command.
(With subtl 0.8.1, it was only required before the join command).

The following error happened when the operator-image was not accicble during the deployment of the broker:
https://qe-jenkins-csb-skynet.cloud.paas.psi.redhat.com/view/Debug/job/Maintenance/job/debug_job/1427/Test-Report/

$ subctl deploy-broker

 • Setting up broker RBAC  ...
 ✓ Setting up broker RBAC
 • Deploying the Submariner operator  ...
 ✗ Deploying the Submariner operator
 ✓ Created operator CRDs
 ✓ Created operator service account and role
 ✓ Updated the privileged SCC
 ✓ Created lighthouse service account and role
 ✓ Updated the privileged SCC
 ✓ Created Lighthouse service accounts and roles
Error deploying the operator: timed out waiting for the condition

subctl version: v0.9

Logs show:

18:04:38 Events:
18:04:38   Type     Reason          Age                  From                                                 Message
18:04:38   ----     ------          ----                 ----                                                 -------
18:04:38   Normal   Scheduled       <unknown>                                                                 Successfully assigned submariner-operator/submariner-operator-8d46d89b7-gtc4n to ip-10-1-190-237.us-east-2.compute.internal
18:04:38   Normal   AddedInterface  10m                  multus                                               Add eth0 [10.133.2.189/23]
18:04:38   Normal   Pulling         9m2s (x4 over 10m)   kubelet, ip-10-1-190-237.us-east-2.compute.internal  Pulling image "registry.redhat.io/rhacm2-tech-preview/submariner-operator-rhel8:v0.9"
18:04:38   Warning  Failed          9m1s (x4 over 10m)   kubelet, ip-10-1-190-237.us-east-2.compute.internal  Failed to pull image "registry.redhat.io/rhacm2-tech-preview/submariner-operator-rhel8:v0.9": rpc error: code = Unknown desc = (Mirrors also failed: [image-registry.openshift-image-registry.svc:5000/submariner-operator/submariner-operator-rhel8:v0.9: Error reading manifest v0.9 in image-registry.openshift-image-registry.svc:5000/submariner-operator/submariner-operator-rhel8: name unknown]): registry.redhat.io/rhacm2-tech-preview/submariner-operator-rhel8:v0.9: Error reading manifest v0.9 in registry.redhat.io/rhacm2-tech-preview/submariner-operator-rhel8: unknown: Not Found
18:04:38   Warning  Failed          9m1s (x4 over 10m)   kubelet, ip-10-1-190-237.us-east-2.compute.internal  Error: ErrImagePull
18:04:38   Normal   BackOff         8m47s (x6 over 10m)  kubelet, ip-10-1-190-237.us-east-2.compute.internal  Back-off pulling image "registry.redhat.io/rhacm2-tech-preview/submariner-operator-rhel8:v0.9"
18:04:38   Warning  Failed          26s (x43 over 10m)   kubelet, ip-10-1-190-237.us-east-2.compute.internal  Error: ImagePullBackOff

[tpantelis]

I'm assuming you ran subctl deploy-broker with --repository and --version set. Those are used to obtain just the operator image. With the join command, we don't explicitly document that custom repo/images must be accessible. We can do that but, not trying to be facetious, shouldn't that be obvious/implied? The commands also take a kubeconfig but we don't explicitly say that it must exist and be available, that's implied.

[Jaanki]

I agree with @tpantelis here. If there is a consensus, can I close this?

[skitt]

I agree too, we shouldn’t need to document this.