Description of problem:

After I triggered the GPU addon uninstall process, the namespace "redhat-nvidia-gpu-addon" is stuck under "terminating" status with the following message:

Some content in the namespace has finalizers remaining:        
    foreground-deletion in 1 resource instances, nvidia-gpu-addon in 1        
    resource instances

All the pods have been deleted and OCM UI reports the Addon as "uninstalled".

The CRs that are blocking the uninstall process are:

• ocp-gpu-addon (NodeFeatureDiscovery)
• nvidia-gpu-addon (GPUAddon)

Prerequisites (if any, like setup, operators/versions):

Install RHODS

Install Nvidia GPU Addon

Steps to Reproduce

1. Go to OCM (https://console.redhat.com/)
2. select your cluster
3. go to Addons section
4. select Nvidia GPU card
5. click on "uninstall" link

Actual results:

"redhat-nvidia-gpu-addon" is stuck under "terminating" status , so uninstall is not completed. 

OCM rerports the addon as uninstalled (it seems there are no checks on the namespace status)

Expected results:

all the resources created by the operator must be deleted when uninstall is triggered

Reproducibility (Always/Intermittent/Only Once):

Reproduced 3/3, 2 different clusters

Build Details:

RHODS v1.13

Nvidia GPU Addon 1.0.0

Workaround:

removing finalizers from the CRs which are blocking the uninstall

Additional info: