NSX Advanced Load Balancer - Part I - Installation with Easy Deploy -

This is the first post in a series on the NSX Advanced Load Balancer. I have never installed the NSX Advanced Load Balancer, but I wanted to test it out in VMC. There is a Fling to help – Easy Deploy for NSX Advanced Load Balancing.

Step 1 – Install the OVA

I downloaded the Fling OVA from the Flings site and deployed it to my vCenter. The installer is a typical next-next-finish OVA deployment. This is a mix of screenshots, some from when I deployed with a static IP, and some from when I deployed with DHCP.

After I deployed with a static IP, I powered up the appliance. Then I of course discovered I had an IP conflict. I redeployed using DHCP on a different segment.

Step 2 – Access the Web Interface

Browse to the IP via https, log in with username admin and the password used at deploy time.

The main deploy screen opens up. I click on first step to add an image.

Step 3 – Add an image

I click on New Image to add a load balancer image

I log into Avi Pulse with My VMware (now called Customer Connect) credentials

I agree to the EULA

The image begins downloading

When the image shows Ready, proceed with a deployment.

Step 4 – Create a Deployment

As of the date of this blog post, the Fling only supports VMware Cloud on AWS. To use it, you must provide an API token. See my blog post on from the SDDC Import/Export series if you need to learn how to generate an API token.

Note: before you create a deployment, you will need 3 network segments:

Controller segment for the AVI controller
SE management segment for the service engines
VIP segment for your virtual IP addresses – these would be the IP that you point DNS to i.e. vmware.com

You will need a 4th network segment if you you want to take advantage of the Fling feature for deploying a sample load balanced application

Backend segment for your backend IP addresses – these are assigned to the nodes. You might also hear this called ‘real IP’. These are the IPs assigned to each server that forms the load balanced application

The interface does not refresh itself, so these segments have to exist in VMC before you start a new deployment.

Click on Deployments, then New Deployment

VMC on AWS is the only choice.

Paste the API token in and click Next

Pick the correct org from the dropdown and click next.

Pick the deployment SDDC from the dropdown and click Next

The controllers are fairly large VMs – if you’re just deploying to play with in a lab, you can skip the cluster and just deploy a single node. Pick the management network segment from the dropdown and click Next

Pick the service engine configuration – leave the default for a simple lab.

Pick a network segment for service engine management. I have DHCP enabled on this segment so SE IPs get autoassigned.

This section can get confusing as there are 2 sets of IP assignments happening from the VIP CIDR. DHCP is assigning IP addresses to the VIP interface of each service engine. The VIP IP start and end IPs are the range assigned to application VIPs. You need to make sure that your DHCP range does not overlap with the range you specify in the VIP pool.