vRA 8 + NSX-T Blog Series Part 7: Create a vRA 8 (Cloud) Blueprint with Existing NSX-T One-Arm Load Balancer
You can create a vRA Cloud blueprint to deploy machines and place them behind an existing NSX-T one-arm load balancer. This method does not create any on-demand tier-1 routers. It uses the existing load balancer and simply adds virtual server, server pool, and application profile (monitor too if you want). This method should work in vRA 8.1 as well, but I used vRA Cloud to create the demo.
Note that vRA Cloud / vRA 8.1 does not fully support NSX-T policy API yet. Full Support for NSX-T policy API is scheduled to be added in vRA 8.2 (but who knows, I’m not a product manager).
If you try to create a blueprint using policy tier-1 router and load balancer, the blueprint deployment will fail and display the following error message:
 [Principal 'admin' with role '[enterprise_admin]' attempts to delete or modify an object of type LoadBalancerService it doesn't own. (createUser=nsx_policy, allowOverwrite=null)]
Demo Product Versions
- VCF 188.8.131.52-15345960 (vSphere 6.7, NSX-V 6.4.6)
- NSX-T 2.5.0
- vRA Cloud
- NSX-T account connected
- Basic infrastructure configured in Cloud Assembly (Projects, Cloud Zones, Flavor Mappings, Image Mappings)
- tier-0 logical router configured
- (non-policy-API) tier-1 logical router configured
- edge cluster configured (edge nodes must be at least medium size)
- Create a NSX-T one-arm load balancer.
- Create a network profile to use an existing load-balancer.
- Create a blueprint with Cloud Agnostic Machine, Cloud Network, and Cloud Load Balancer.
- Create inputs in the blueprint to customize the machine name.
- Create multiple load balancers and apply a constraint tag to deploy load balancing services on a specific load balancer.
Demo / Example
Create NSX-T One-Arm Load Balancer
- In the NSX-T UI, go to “Advanced Networking & Security” > “Networking” > “Load Balancers”.
- Click “+” to create a new load balancer.
- Click on the load balancer you’ve just created and click “EDIT” next to “Attachment”.
- Select a Tier-1 logical router to provide load balancing services. The tier-1 router cannot be a policy-API router. Create a non-policy-API tier-1 router under “Advanced Networking & Security” > “Networking” > “Routers” if you do not have one.
- You don’t need to configure virtual server or server pools at this time. They will be created by the blueprint.
Create Network Profile
- In vRA Cloud (or vRA 8) Cloud Assmebly, go to “Infrastructure” > “Network Profiles” (under Configure) and click “+ NEW NETWORK PROFILE”. (or you can choose to edit an existing network profile).
- Choose an account/region and give the profile a name.
- Go to “Networks” tab and configure existing networks or on-demand networks. In this demo, I’ll be using an existing NSX-T network. Refer to previous vRA 8 + NSX-T blog series for more details on how to configure existing or on-demand networks.
- When using an existing NSX-T load balancer in a blueprint, you do not have to select tier-0 logical router or edge cluster under “Network Policies” tab.
- Go to “Load Balancers” tab, click “+ ADD LOAD BALANCER” and choose the load balancer created in previous steps.
Create and Configure Blueprint
- Go to “Blueprints” and Click “+ NEW” to create a new blueprint.
- Give a name to the blueprint and choose a project.
- Drag on a Cloud Agnostic Machine, Cloud Agnostic Network, and Cloud Agnostic Load Balancer onto the canvas. You can use environment-specific resources as well if you’d like.
- Connect the Machine and the Load Balancer to the Network on the canvas. Also connect the Machine to the Load Balancer.
- On the right side in the YAML file, choose an image and size for the machine. Add
countproperty to indicate how many machines you want to deploy.
- network:, add the line
assignment: staticto give a static IP address to the machine.
- Since I’ll be placing the machines on an existing NSX-T network, I want to make sure it says
networkType, I also add the line
constraints:then another line
- tag:to choose the existing network I want to use.
- Configure the load balancer.
routesproperty are required.
healthCheckConfigurationis optional - if you provide details, custom monitor will be created. Otherwise, only the virtual server, server pool, and application profile will be created.
- Click “TEST”.
- Click “DEPLOY” to create a new deployment.
- Give it a deployment name, choose “Current Draft”, the click “DEPLOY”.
- Go to “Deployments” tab in Cloud Assembly and check that the deployment is completed successfully.
- Now log into NSX-T UI and go to “Advanced Networking & Security” > “Networking” > “Load Balancers” and you’ll see that the existing load balancer now has a virtual server.
- If you click on the virtual server, you can see the port and protocol as you’ve defined in the blueprint.
- If you click the server pool and look at the pool members, you will see the machines that have been created by the blueprint.
- If you click the profiles, you can see application profile that has been created.
- Since this example had
healthCheckConfigurationproperty specified in the blueprint, you can see custom monitor that has been created.
Note that in my example, the virtual server pool operational state is down because I did not open the appropriate port on the machines to communicate. Assuming your servers are configured correctly, the load balancer should work as intended.
Demo / Example Blueprint YAML File
(Optional) Multiple Load Balancers in Network Profile
If you have multiple load balancers in a network profile, you can put tags on each load balancer then specify the tag in a blueprint to deploy the load balancing services to a specific load balancer.
- Go to the network profile and add tags to each load balancer.
- Add a constraint tag in the blueprint to indicate which load balancer you want to use.
- Once you deploy this blueprint, you will see in the request details / provisioning diagram that the constraint tag is applied.
- After the deployment is complete, go to the NSX-T UI and you’ll see that the load balancing services have been deployed to the desired load balancer.
Using Constraint Tag to Indicate Load Balancer Example Blueprint YAML File