Recommended Audience: Anybody with an Infrastructure, networking, vSphere, and NSX -V background.
What to expect from this post: Prepping the NSX-T environment for configuring “workload management” in vSphere7 with Kubernetes.
NSX-T Implementation Order > Create a Transport Node Profile > Configure Host Transport Nodes using the Transport Node Profile (ESXi Host Preparation)
Here are the high-level implementation steps that we’re going to follow to deploy and configure NSX-T Manager, so we can make it ready for configuring “Workload Management” on vSphere 7.0 to enable Tanzu Kubernetes Grid.
Create a Transport Node Profile
The transport node profile is a template with all the necessary information such as transport zones, uplink profiles, N-VDS or VDS, ESXi TEP IP address details, etc. that can be applied against an existing vSphere cluster or clusters. Since this is a template, it can be repetitively used across multiple clusters and provide consistency.
Click on System (top-menu) > expand Fabric (left-menu) > click on Profiles (left-menu) > click on Transport Node Profiles at the top > add a new transport node profile by clicking on “+ADD” button > provide a name > Under New Node Switch, select > Type: VDS > Mode: Standard (All hosts) > Select the vCenter Server > Select the VDS Switch on which the cluster of transport nodes exist > select the overlay transport zone created in step03 > select the uplink profile created in step04 > select Use IP Pool under IP Assignment > select the IP pool (ESXi TEP) created in step03 > under Team Policy Switch Mapping, select Uplink 1 for VDS Uplinks and click ADD as shown in the screenshots below.
Configure Host Transport Nodes (ESXi Host Preparation)
Now, that the Transport Node Profile is created, we can prepare the transport nodes (ESXi hosts) by either selecting individually or by selecting an entire cluster managed by a vCenter Server.
Click on System (top-menu) > expand Fabric (left-menu) > click on Nodes (left-menu) > click on Host Transport Nodes at the top > under Managed by, select vc-m01 > now, select the cluster that you would like to attach to the transport node profile and prepare > click on “CONFIGURE NSX”.
On the NSX Installation window, select zc_TNP01 for Transport Node Profile and click APPLY as show in the screenshots below.
Now, you should see that the ESXi hosts are getting prepared
Make sure that the NSX configuration is successful and the node status is up for all the ESXi hosts of the cluster.
Now, SSH into ESXi01 host, log in as root and run the following commands:
# esxcli network ip interface ipv4 get > you should see a table of vmk interfaces, especially should see vmk10 and vmk50 with respective IP addresses > you’ll also realize that the IP address assigned
# vmkping ++netstack=vxlan -I vmk10 -s 1400 -d <ESXi02.zerocloud.local>
# vmkping ++netstack=vxlan -I vmk10 -s 1572 -d <ESXi02.zerocloud.local>
# vmkping ++netstack=vxlan -I vmk10 -s 1573 -d <ESXi02.zerocloud.local>
This will determine that the geneve packet sizes are allowed up to 1572 packet size, which is the default. This is the reason, we bumped up the vSphere VDS MTU size to 1600.
That concludes this post and let’s move onto part06.