NSX-T Lab: Transport Node Profile

Intro

Welcome to Part 8 of the NSX-T Lab Series. In the previous post, we created the uplink profile for our compute nodes, so now we can start building our Transport nodes.
So why is step 8 a profile and step 9 configuring the transport node? 
Well let me explain.

What is a Transport Node Profile and why do I care?

OK so we’re going to create a profile for our transport nodes, why not just configure the nodes? Why should we use a profile?
Well remember we added a compute manager the vCenter to NSX, the reason for that was to allow ease of administration and configuration for multiple systems at once and this is where it comes into play.
We could simply go to each standalone host and configure it as a transport node and that will work but it’s not very practical for a production multi host environment and it is also error prone.

So a better approach is to use a Transport Node profile which contains all the settings we need to configure the nodes. We then simply attach the cluster to the profile and configure NSX on all hosts in the cluster at the same time. This also means that new hosts will also be added to the profile when they are added to the cluster.

However if you need to have some hosts in the cluster configured and some not then you can still deploy the nodes manually one by one.
This is where NSX-T and NSX-V differ.
When preparing the hosts for NSX-V it is done on the cluster level and as such all hosts in the cluster must be prepared.
With NSX-T we have the option to prepare all or only some hosts.
I would still say that its generally a good idea to prepare all hosts in a cluster, if you need hosts to not be part of NSX-T then best bet is to move them to a different cluster.

What I mean by prepare is to deploy NSX VIB’s onto the hosts.

The build

I can’t keep saying “from the NSX Console” since all NSX configuration has to be done from the web console so lets now assume you know where to be to start any configuration work 😉

Under ‘System’ ‘Fabric’ ‘Profiles’ Select the Transport Node Profiles tab and go ahead and hit ‘+ Add’

Give the Profile a name and optional Description.
We now need to add the Transport Zone/s that the Nodes will be part of.
For my lab the compute nodes are only a part of the Overlay Transport Zone so I add the TZ-Overlay.
You also have the option from this screen to create a new Transport Zone if you need to.
Don’t click add just yet as we need to configure the N-VDS first, so click on the N-VDS tab.

From the options select the N-VDS, there is likely only one option to pick from here since the Transport Zone only has one N-VDS assigned to it.

For the NIOC profile we’re just going to use the default.

Change the Uplink Profile to the one we created in the last step.

LLDP Profile can be either Send Packet Enabled or Disabled.
When LLDP is enabled the host hypervisor advertises it’s network information to the peering transport nodes.
For NSX Edge nodes this setting needs to be disabled.

Under IP Assignment select use IP Pool
This adds a new option of IP Pool, go ahead and select the IP Pool we created earlier. If you didn’t create one then you can do it from here.

Under Physical NICs enter the vmnic number, this NIC needs to be unassigned on the host otherwise it won’t work, for my lab the first unused NIC is vmnic1.
As the uplink profile contains two uplinks we need to add a second one so click on Add PNIC and set it to the next available NIC, for me thats vmnic2. Now we can click on Add.

We now have a Transport Node profile ready to be used in the next step.

Finally we are ready to actually install NSX-T on our hosts.
In the next post I’ll cover how to deploy using the Transport node profile and via a standalone host and you will see why we created this profile.

NSX-T Lab Part:9 NSX-T Host Transport Nodes

Leave a Reply

Your email address will not be published.