CVO on GCP and network interfaces

[pcmssysadmin]

Hi all bit of an obscure question about deploying CVO on GCP.

Network interfaces are set at creation of the compute instance is it possible to add more interfaces.

To try and explain why this has come about the question is coming from the architect designing our cloud systems and he is wanting to do something in cloud similar to what we do on-prem with vlans to secure/split up traffic but using multiple nics in GCP.

As an example his requirement would be to have a project in GCP with multiple dev environments running as separate compute instances each with a separate VPC. He would like to know if we can deploy the CVO workspace to then have a network interface in each of the dev VPC networks rather than just having one interface and using network rules to direct the  traffic from the device VPC networks to the controller VPC network. Looking on a deployed controller all of this should be possible as the commands are from what we can see all there we just need to know if we can deploy the controller/s with extra nics.

[elementx]

It sounds like he wants multiple SVMs which presently doesn't seem doable in default configuration:

https://docs.netapp.com/us-en/occm/task_managing_svms.html

[rnicholl · Accepted solution]

Adding NICs after deployment is not supported in CVO in GCP.

[pcmssysadmin]

Yep one of those things you don't find out until you go to try to do it 🙂

What I have tried which was unsuccessful so far was to create a clone of the appliance and add on the extra hardware nics when detailing the new instance but unfortunately while cloud manage can see the clone is there it wont boot up.