Part 4 of the journey takes us through deploying NSX-T on our ESXi Hosts. This will be a multi-step process and has some complexity but this piece can be a lot of fun. The work we do here will have a ripple effect as we move forward so it is important to know your use case and the best design to support it.
Let the fun begin:
Step 1. Create a Uplink Profile to match your pNIC configuration
Go to System>Fabric>Profiles>Uplink Profiles>Add
Enter in a name that
specifies how many ports. IE if you are
going to assign 3 x pNICs then I recommend a name like Uplink-Hosts-3Uplinks
- Failover Order: An active uplink is specified along with an optional list of standby uplinks. If the active uplink fails, the next uplink in the standby list replaces the active uplink. No actual load balancing is performed with this option.
- Load Balanced Source: A list of active uplinks is specified, and each interface on the transport node is pinned to one active uplink based on the Source Port ID. This configuration allows use of several active uplinks at the same time.
- Load Balanced Source Mac: This option determines the uplink based on the source VM’s MAC address.
- The Load Balanced Source and Load Balanced Source Mac teaming policies do not allow the configuration of standby uplinks.
Since I have plenty of NICs in my environment I built my new profile with a total of 4 x pNICs.
If you are using LAGs here is where you will enter the information.
You can click edit next to failover order if you want to change the Teaming Policy. IE my recommended selection of Load Balanced Source.
Finally, you will want to enter you Transport VLAN and MTU size.
Step 2. Create
First, let’s create a TZ for our overlay network. I recommend calling each TZ what it will be a TZ for in this case, TZ-Overlay. Also, we need to create a N-VDS and each name must be unique so something like TZ-Overlay-Prod
When done, click add.
Do the same for an Edge TZ only this time select traffic type as VLAN.
Once again I would put VLAN in the name.
I’m also going to have additional TZs for my Edge Uplink1 & 2. This is specific for my environment but isn’t a bad design for more network isolation, HA, and flexibility with BGP Peering.
Finally, wait till they are all in a status of up and we can move onto creating IP Pools
Step 3. Create IP Pools
You create an IP pool for assigning IP addresses to the NSX transport nodes
Networking>IP Address Management>IP Address Pools
Provide a name, I recommend something that allows you to know what the TEP Pool is associated with. Something like, Overlay-TEP-Pool
Then click Set
Put in the range of IP’s xx.xx.xx.xx-xx.xx.xx.xx
Put in the associated CIDR
Put in the Gateway IP
If you are doing a collapsed edge/compute leveraging a N-VDS like I am, then repeat this process for your VLAN TEP as well. Your Edge and Overlay must be on separate VLANs for this architecture.
Look for the status of Up for both your new IP Pools
Step 4. Setup
Drop down; Managed
by: select your vCenter
Select the cluster
you want to deploy NSX-T on and click Configure NSX
Create new Transport Node Profile
Create new Transport Node Profile
Provide a name. I recommend ESXi Profile in case you ever decide to run KVM.
Select all of your TZ’s and move it to the right.
Select N-VDS at the top
Select the N-VDS we created earlier when we created our TZ
Select a NIOC Profile: I keep mine default
Uplink Profile: Select the Uplink Profile we created earlier
IP Assignment: Use IP Pool
IP Pool: Select the TEP IP Pool we created earlier for our Overlay
Time to add our VLAN N-VDS to the profile.
I keep everything default for this.
Should everything go
to plan 🙂 We should have under NSX
Configuration: Configured, Configuration State: Success, Node status: Up,
Transport Zones: the name of your Overlay TZ
There are a lot of moving parts here. As you can see applying NSX-T to your ESXi hosts is not a single click of a button. However, this is not a process that takes a day or two but only a couple of hours. The planning for this piece should take days 🙂 This is a key part of deploying your NSX-T infrastructure. Next in the process will be deploying the Edge VMs and combining them into an Edge Cluster.