Upgrade Antrea-NSX Interworking Deployment in an Antrea Container Cluster

If you have upgraded the VMware Container Networking version to a new release, the Antrea-NSX interworking deployment in your registered Antrea container cluster must be upgraded together.

Prerequisites

The VMware Container Networking version in your Antrea container cluster, which is registered to NSX, must be upgraded successfully.
Complete only the following prerequisite steps that are mentioned in Prerequisites for Registering an Antrea Container Cluster to NSX-T Data Center. Skip the other prerequisite steps at this link.
- Determine the Antrea version from the Kubernetes cluster.
- Download the Antrea-NSX interworking zip file.
- Import the container images to container registry.
  Make sure to edit the interworking.yaml and deregisterjob.yaml files, and update the image URLs to the imported image, or to the online image URL.

Procedure

Run the following kubectl commands to delete the register job and the antrea-interworking deployment.
```
kubectl delete job register -n vmware-system-antrea --ignore-not-found
kubectl delete deployment antrea-interworking -n vmware-system-antrea --ignore-not-found
```
Note: Do not delete the vmware-system-antrea namespace, configmaps, and secrets in this namespace.
Run the following kubectl command to submit the interworking.yaml file to the Kubernetes API server, and trigger the upgrade.
```
kubectl apply -f interworking.yaml
```
Note: Only the interworking.yaml must be submitted. The bootstrap-config.yaml file is not required for the upgrade process.
Run the following kubectl command to list all the Pods in the vmware-system-antrea namespace.
```
kubectl get pods -o wide -n vmware-system-antrea
```
Observe that the status of the register-xxx Pod is Running. Because the Antrea container cluster is already registered to NSX, the register-xxx Pod skips the registration process and the status soon changes to Completed. The old interworking-yyy Pod status changes to Terminating, and the new interworking-zzz Pod status changes to Running.
When the new interworking-zzz Pod status is Running, and the ready containers are 4/4, there is no need to restart the containers, and the upgrade is successful.
```
#Example output:

NAME                            READY   STATUS    RESTARTS   AGE   IP              NODE              NOMINATED NODE   READINESS GATES
interworking-7764988ddd-wnvcg   4/4     Running   0          29s   192.168.x.y     example-node-10   <none>           <none>
```

Run the following kubectl command to verify that the new interworking-zzz Pod is using the new image and the image URLs are the same as expected.

Make sure to replace the "interworking-7764988ddd-wnvcg" Pod name with the actual Pod name that you see in the output of the kubectl get pods command of the previous step.

kubectl get pods -o yaml interworking-7764988ddd-wnvcg -n vmware-system-antrea | grep image:

# Example output:
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0
    image: vmware.io/antrea/interworking:0.11.0