Multizone clusters with GPU_NODE_MIN_SIZE=1 must have a GPU quota of at least 2.

[willgraf]

Describe the bug
Deploying a cluster in multiple zones requires a GPU quota equal to the number of zones if GPU_NODE_MIN_SIZE is at least 1. If the GPU quota is only 1, a multizone cluster will fail with the error:

ERROR: (gcloud.container.node-pools.create) ResponseError: code=403, message=Insufficient regional quota to satisfy request: resource "PREEMPTIBLE_NVIDIA_V100_GPUS": request requires '2.0' and is short '1.0'. project has a quota of '1.0' with '1.0' available. View and manage quotas at https://console.cloud.google.com/iam-admin/quotas?usage=USED.

To Reproduce
Steps to reproduce the behavior:

1. During cluster configuration, select a GPU type with a quota of 1.
2. Select a minimum GPU node count of at least 1.
3. Select a multizone cluster configuration.
4. Create the cluster.
5. See the error in the failed cluster creation output.

Expected behavior
Ideally, the cluster will not allow the user to start creating a cluster with a bad configuration.

[willgraf]

The solution to this bug is tracked in #335, though I will leave this issue open as a documented warning until it is resolved.