Description of problem:

The assisted-service and assisted-image-service pods do not include imagePullSecrets in their spec and end up in ImagePullBackOff state when MCE is deployed on a non-OpenShift k8s cluster.

Version-Release number of selected component (if applicable):

MCE 2.9

How reproducible:

Behavior is consistent when deploying on non-OCP cluster.

Steps to Reproduce:

1. Deploy MCE on a non-OpenShift Kubernetes cluster (e.g. by installing the Helm chart at https://github.com/Azure/ARO-HCP/tree/main/acm/deploy/helm/multicluster-engine)
2. Define AgentServiceConfig CR to enable infrastructure-operator

Actual results:

assisted-service and assisted-image-service are in ImagePullBackoff state...

$ kubectl get pods -n multicluster-engine | grep assisted
assisted-image-service-0                                 0/1     ImagePullBackOff    0               85s
assisted-service-db467869c-6mkbx                         0/2     ImagePullBackOff    1 (30s ago)     85s 

Expected results:

Pods are running.

Additional info:

The issue can be worked around by patching the service accounts with the pull secret...

kubectl patch sa assisted-service -p '{"imagePullSecrets": [{"name": "open-cluster-management-image-pull-credentials"}]}' -n multicluster-engine
kubectl patch sa assisted-image-service -p '{"imagePullSecrets": [{"name": "open-cluster-management-image-pull-credentials"}]}' -n multicluster-engine 

However, this doesn't lend itself to automated deployments.