VMware vSphere 8 Tanzu Deployment using HA proxy and VDS

Requirements for Tanzu Deployment on vSphere: –

1. vCenter min version – v7.0U1
2. Officially recommend 3 ESXi hosts
3. HA and DRS enabled
4. Three networks (VDS or NSX-T)
5. Ha proxy or AVi Load Balancer
6. Storage
7. Tanzu License (You can try with 60 Days Evaluation License )

My Tanzu Deployment Environment: –

1. vSphere – 8.0
2. Single ESXi Host   – ESXi 8.0
3. HA and DRA enabled
4. VDS Network     – Three Ports Configured – HAProxy-LB-92, TKG-Frontend-93, TKG-WKLD-91
5. HA Proxy

Networks CIDRs: –

VDS Port Name	CIDR
HAProxy-LB-92	172.92.0.0/24
TKG-WKLD-91	172.91.0.0/24
TKG-Frontend-93	172.93.0.0/24

IP Configuration: –

IP	Use	Comment
172.92.0.5	HA Proxy VM IP
172.92.0.20 – 172.92.0.25	Supervisor VM’s IP	3-node Supervisor cluster deployed as part of workload management cluster deployment.
172.93.0.150-172.92.0.180	VIP	Each Tanzu cluster needs One VIP. Using VIP only, we can access Cluster.
172.91.0.20-172.91.0.100	Workload	For Deploying Workload Kubernetes clusters
192.168.5.150	DNS	DNS name – kdinesh.in
0.in.pool.ntp.org	NTP

Deployment Steps: –

1. Configure Storage policy and Tags
2. Configure Tanzu Content Library
3. Downloading Ha proxy Ova and deploying Ha proxy OAV
4. Enabling Tanzu Workload Management
5. Creating Namespace and deploying Tanzu workload Kubernetes cluster

1. Configure Storage Policy and Tags

Login to vSphere > Tags & Custom Attributes

Categories > new

Category Name and Select all objects > Create

Tags > new

Tag name > Select Category > Create

Select datastores (Which datastore want to use for tkg deployments)> Right Click on datastore > Tags & Custom Attributes

Select the newly created tag > Assign

Check tag is assigned to the datastore or not

Select the policy and profiles option

Select VM Storage and Policies > Create

Enter Policy Name > Next

Select Enable Tag Based Placement rules option > Next

Browse previously created tag > Next

Check tagged Datastore is showing in the compatible list or not > next

Finish

• Configure Tanzu Content Library

Select the Content Libraries option

Create

Enter Name > Next

Select Subscribed content library   – URL – https://wp-content.vmware.com/v2/latest/lib.json  >

Choose the immediately option (automatically, it will download all Tanzu templates) > Next.

Yes

Next

Select datastore (where you want to store Tanzu templates/ova – approx. size – 125+ GB) > Next.

Finish

Downloading Tanzu OVA/templates take a few minutes, based on your internet speed.

You can check the status – Total of 31 Tanzu templates,125.27Gb.

• Downloading Ha proxy Ova and Deploying HA Proxy LB

You can download Ha proxy OAV from – https://github.com/haproxytech/vmware-haproxy#download

After downloading Deploy Ha proxy OVA on vSphere

Right-click on host or cluster > Deploy OVF Template

Upload Local file > select Downloaded Ha proxy Ova File > Next

Enter Virtual machine name > Select Location >Next

Select Compute Resource > Next

Next

Accept License > Next

Select Frontend Network > Next

Select datastore (Where you want to deploy VM) > next

Select management, Workload, and Frontend Network > Next

Enter Required details (Host name, Mgmt Ip, gateway, Workload IP, Gateway, Frontend IP, VIP.etc.) > Next

Finish

  Power on Ha proxy VM

After power on > Login to Ha proxy and copy ca.crt

Cmd –   cat /etc/haproxy/ca.crt

• Tanzu Workload Management Deployment

Select Workload Management Option > get started

Select vCenter > select vDS > Next

Select cluster Deployment >Enter Supervisor name > Select cluster > next

Select Previously created Storage policy > Next

Enter required details (lb name, IP, username and password (during ha proxy deployment configured), VIP Range, paste previously copied ha proxy certificate on tls certificate section > next

Select network based on network type (In my case, it’s static)>Specify stating IP (Usually it will use 4 IPs for supervisor nodes)  ..etc. details > next

Select Workload Network > enter IP pool…. etc. > Next

Select previously created Tanzu content library > Next

Select Control plane size based on your requirement > Finish

Cluster Provision will take a few minutes. Once Cluster is up, you can check the status

You can check on the vSphere three-node supervisor cluster created

• Creating Namespace and deploying workload Kubernetes cluster

Select Namespaces > Create Namespace (Once Supervisor cluster deployment is completed, you need to create a namespace, in namespaces, you can deploy multiple workload clusters based on your requirements) > Enter Cluster name > Select workload network > Create

Once the namespace is configured > you need to configure permissions, storage, VM classes, and content library to the namespace.> Select ADD permission

Select Identity source, Group/user, Role (Owner, can edit or Can view)> Ok

Next, Select Storage policy (Previous steps created policy, select policy) > Ok

Select ADD VM Class > select based on your requirements VM classes (For deploying the workload cluster, you can choose these VM classes > Ok

Next, Add content lib > Select previously   created lib > OK

Check whether all are configured or not

Open Browser and Enter Supervisor Cluster VIP   > Based on your OS, you can download CLI Plugin (for access the deployed cluster and deploying new workload clusters)

Downloaded and extracted CLI plugin on windows > you can see two files, kubectl and kubectl-vSphere files > Open CMD and go to the path

Connect to vSphere Cluster

CMD – kubectl vsphere login –server=” Supervisor cluster VIP” –insecure-skip-tls-verify –vsphere-username “vsphere username “

You switch back to the previously configured cluster

CMD – Kubectl config use-context clustername

Created Workload deployment yaml file with workload specifications (one master, one worker node with small class)– wkld1.yaml (based on your requirement, you can create)

Run wkld1.yaml file (kubectl apply -f filename). It will provision workload cluster based on the file.

You can check recent tasks on vsphere, Workload deployment stated. It will take a few minutes to provision the cluster.

Once deployment is completed you can check the status of running or fail

You can also check through CLI

You can also check in vSphere two Node workload cluster configured (because I specified in yaml file, 2 nodes cluster)

Similarly, way you can deploy multiple clusters based on requirements and once the cluster Is ready you can deploy pods on workload clusters.