mist_device_gateway_cluster not removing devices on a destroy

[zbuchheit]

Describe the bug

When attempting to destroy a gateway_cluster it seems the devices are being left in the cluster and causes a subsequent apply to fail with the devices still part of the cluster

To Reproduce

Steps to reproduce the behavior:
TODO

Expected behavior

I would expect for the devices to be removed from the cluster and running an apply to re-create the cluster successfully

Error Message

If applicable, add screenshots or paste the Terraform error message to help explain your problem.

Software

• OS: [e.g. Windows]
• Terraform Version [e.g. v1.9.8]
• Provider Version [e.g. v0.2.6]

Additional context

Add any other context about the problem here.

[zbuchheit]

added for tracking purposes and comms, thanks @tmunzer

[tmunzer]

Hi @zbuchheit,

The only way I found to replicate the issue was to remove the devices part of the cluster from the org_inventory resource without destroying the gateway_cluster (or unassign them from the site).
To be honest I'm not sure I'll be able to find a way to improve this.

However, I'm bringing changes to the device_gateway_cluster resource to not raise an error when a planned cluster must be created but already exists in Mist. The providing is now importing the existing cluster in the state if the node MAC addresses in Mist are matching the planned nodes MAC Address.

[zbuchheit]

hmm makes sense. I think that improvement would be a welcomed one and help fix a potential broken scenario