danielgarner64
Contributor
Contributor

VRA Cloud Template with 2 NSX-T Networks

Jump to solution

VRA Version 8.4, Vsphere version 7.x, NSX-T 3.1

I'm attempting to deploy a simple VRA Cloud Template with two Nics (both NSX-T networks).
The problem is when I deploy the template both the NICs will have the same IP range and they shouldn't. (One Network Profile clobbers the others IP range).

Here is some detail to make it clear what is happening.

The template looks something like this.

formatVersion: 1
inputs: {}
resources:
  Nsxone:
    type: Cloud.NSX.Network
    properties:
      networkType: private
      constraints:
        - tag: 'otherTagB:hard'
  NsxTwo:
    type: Cloud.NSX.Network
    properties:
      networkType: private
      constraints:
        - tag: 'customTag:A:hard'
  VirtualMachineNode:
    type: Cloud.vSphere.Machine
    properties:
      flavor: Medium
      image: Win10Ansible
      networks:
        - network: '${resource.NsxTwo.id}'
          deviceIndex: 0
        - network: '${resource.Nsxone.id}'
          deviceIndex: 1

 Within the provisioning diagram I can see that each network is independently allocated to each tagged resource network profile.

The details of the two profiles are listed below but effectively they are identical except they have different CIDRs:

Network Profile with Tag otherTag:B

Infastructure -> Network Profiles -> Summary

- Name: NetworkB

Capability tags: otherTag:B

Infastructure -> Network Profiles -> Network Policies
Isolation Policy:

- Transport Zone: overlay

- Tier-0 Logical Router: gateway

- Edge Cluster: Edge-Cluster

IP Address managment:

source: internal

CIDR: 5.5.0.0/16

Subnet Size: /24

IP Range assignment: static and DHCP

 

Network Profile with TagcustomTag:A

Infastructure -> Network Profiles -> Summary

- Name: NetworkA

Capability tags: TagcustomTag:A

Infastructure -> Network Profiles -> Network Policies
Isolation Policy:

- Transport Zone: overlay

- Tier-0 Logical Router: gateway

- Edge Cluster: Edge-Cluster

IP Address managment:

source: internal

CIDR: 10.22.0.0/16

Subnet Size: /24

IP Range assignment: static and DHCP

 

 

After the deployment completes the two NSX networks will have CIDRs from the same range.

For example one network will be NsxOne will have CIDR: 10.22.1.0/24 and NsxTwo will have CIDR 10.22.2.0/24

OR

For example one network will be NsxOne will have CIDR: 5.5.1.0/24 and NsxTwo will have CIDR 5.5.2.0/24


The expected result from deploying the network should be:

For example one network will be NsxOne will have CIDR: 10.22.1.0/24 and NsxTwo will have CIDR 5.5.1.0/24

 

I have also tested with no virtual machine present and I get the expected result. i.e NsxOne will have CIDR: 10.22.1.0/24 and NsxTwo will have CIDR 5.5.1.0/24

I have also tested with two virtual machines one connected to each network and I got the expected result. i.e NsxOne will have CIDR: 10.22.1.0/24 and NsxTwo will have CIDR 5.5.1.0/24

When I tested it with two virtual machines one connected to both networks and one connected to one network I did not get the correct result I got something like NsxOne will have CIDR: 10.22.1.0/24 and NsxTwo will have CIDR 10.22.2.0/24

For those who prefer to see the issue below are two images of the above two scenarios I just described:
No Issue with No Dual NIC

Expected result with no Dual NICExpected result with no Dual NIC

Dual Nic With Issue

Incorrect result with Dual NicIncorrect result with Dual Nic

 

 

0 Kudos
1 Solution

Accepted Solutions
danielgarner64
Contributor
Contributor

So after contacting VMware through VMware support this is a but that exists in VRA 8.4.0.

It is fixed in VRA 8.4.2, the only fix is to update VRA.

View solution in original post

1 Reply
danielgarner64
Contributor
Contributor

So after contacting VMware through VMware support this is a but that exists in VRA 8.4.0.

It is fixed in VRA 8.4.2, the only fix is to update VRA.