cattle-node-agent-windows fails to deploy on a windows cluster having nodes from different VPC

[sowmyav27]

What kind of request is this (question/bug/enhancement/feature request): bug

Steps to reproduce (least amount of steps as possible):

1. Deploy a windows cluster
2. Cluster spec - 1 AWS linux node from Ohio region, 1 AWS windows node from Oregon region and 1 windows GCE node.
3. Register and enter the public IP addresses of the nodes while registering.
4. The cluster comes up Active

Expected Result: All workloads under cluster --> System Project should be deployed successfully.

Actual Result: cattle-node-agent-windows fails to deploy on the windows workloads

Other details that may be helpful:

Environment information

• Rancher version (rancher/rancher/rancher/server image tag or shown bottom left in the UI): master-head
• Installation option (single install/HA): single

Cluster information

• Cluster type (Hosted/Infrastructure Provider/Custom/Imported): custom
• Kubernetes version (use kubectl version):

1.15.3

• Docker version (use docker version):

19.3.1

[thxCode]

AFAIK, you could not operate the Pods which scheduled on the GCE node via kubectl, even in a pure Linux cluster. For now, we could not support host communicated with Flannel VxLan on different VPCs.

[sowmyav27]

Verified on a linux cluster on master-head-- Linux cluster has 2 nodes, Both from AWS but different regions Exec log and logs DO NOT work on the workload

[sirredbeard]

@rosskirkpat Can you see if you can reproduce this? If so, how would you recommend we resolve this issue?

[rosskirkpat]

@rosskirkpat Can you see if you can reproduce this? If so, how would you recommend we resolve this issue?

Sure, I'll take a swing at attempting to reproduce this