OCP: Assisted Installation

In this optional lab we will go through Assisted Installation feature that Red Hat Console offers.

Assisted Installer is a GUI driven way of deploying OCP clusters.

You would use Assisted Installer clusters for to deploy OCP clusters quickly (e.g. before a POC) and for personal testing purposes.

Single node OCP cluster implementation is also available with Assisted Installer. We will only look at deploying multi-node clusters in this lab.

note

Estimated time to complete this lab is 60 minutes.

caution

If you are doing this section of the lab using a Single Node HPOC (SPOC) there will not be enough resources to deploy a second OCP cluster.

Remove existing OCP clusters that you don't require anymore.

If you were using NCM (Calm) - Please delete the other OCP cluster(s) from Calm > Actions > Delete before proceeding with this lab.

Pre-requisites for Assisted Installation

Administrator finds and reserves API and Ingress IP addresses for OCP cluster
Administrator creates the OCP Master and Worker VMs
Adminsitrator provisions compute, networking and storage associated with the OCP Master and Worker VMs
Access to Red Hat Console to use Assisted Installer and install a OCP cluster

High Level Overview of Assisted Installation

Assisted Installer does the following:

Provides RHCOS and OCP installation binaries in a CD-ROM ISO file
Once the OCP VMs (Master and Worker) nodes are booted with this CD-ROM ISO file, using the public key the VMs connect to Red Hat Console
VMs show in Assisted Installer page and the administrator (you) begins the installation
Red Hat Console will manage and monitor the installation process from start to end
Upon successfull installatino of a OCP cluster the following will be provided:
- KUBECONFIG file for oc access
- DNS entries for OCP Cluster Manager access
Assisted installer will choose one of the Master VMs to serve the Bootstrap role

info

Assisted Installers in GA as of July 2022 General Availability .

The OCP clusters deployed using Assisted Installers can be used for production, testing and development purposes owing to pre-requisites conformity.

At a very high level, we will accomplish the following to get a OCP cluster deployed using Assisted Installer:

Find and reserve two IPs in our AHV network for OCP cluster's API and Ingress endpoints
Provision OCP Cluster in Red Hat Console and Generate CD-ROM ISO URL
Provision OCP Infrastructure - Create Master and Worker VMs in your AHV HPOC/SPOC cluster using Terraform infrastructure as code
Install OCP Cluster in Red Hat Console

OCP Assisted Installation

Pre-requisites

AHV Networking

We will first find two IPs for OCP api and apps ingress endpoints in our network and add it to the Primary IPAM network blacklist.

Find the CIDR range for your Primary IPAM network either from RX or from your instrutor
CIDR example for your Nutanix cluster
```
10.38.18.192/26
```
Logon to your UserXX-LinuxToolsVM (deployed on the HPOC cluster) using Mac Terminal or Windows Putty
Username: root
Password: nutanix default

Prepare the UserXX-LinuxToolsVM

yum update -y
yum install -y bind-utils nmap

Find two unused static IP addresses

nmap -v -sn  <your CIDR>

Sample command
nmap -v -sn 10.38.18.192/26

Sample output - choose the first two consecutive IPs
Nmap scan report for 10.38.18.219 [host down] 
Nmap scan report for 10.38.18.220 [host down]
Nmap scan report for 10.38.18.221
Host is up (-0.098s latency).

Logon to any CVM in your cluster and execute the following to add chosen static IPs to the Primary IPAM network
- Username: nutanix
- Password: your cluster password
```
acli net.add_to_ip_blacklist <your-ipam-ahv-network> ip_list=10.38.18.219,10.38.18.220
```
Sample command
```
acli net.add_to_ip_blacklist Primary ip_list=10.38.18.219,10.38.18.220
```

Add DNS Records

In this section we will add PC, API and APPS Ingress DNS records for lookup by OCP IPI installer.

Your OCP cluster's name becomes a subdomain in your DNS zone ntnxlab.local. All OCP cluster related lookups are located within subdomain.

Main domain - ntnxlab.local (this gets created with your HPOC reservation)
- Sub domain - ocpuserXX.ntnxlab.local (e.g. ocpuser01, ocpuser02, etc, is your OCP cluster's name)

Logon to the AutoAD windows VM
Username: administrator
Password: default
We will add the following entries to DNS server using the two consecutive IPs you found in the previous section
Use your assigned HPOC cluster's IP Addresses
The IP addresses in the following commands are used as an example. You should use IP address details that belong to your HPOC cluster.
For information on locating your cluster IP see Lookup website.
```
10.38.18.219   api.ocpuserXX.ntnxlab.local
10.38.18.220   *.apps.ocpuserXX.ntnxlab.local
10.38.18.201   pc.ntnxlab.local
```

Open PowerShell as Administrator and create the two A records

Template commands
Sample commands

Add the API A record - use your own subdomain

Add-DnsServerResourceRecordA -Name api.<ocpuserXX> -IPv4Address <your API IP> -ZoneName ntnxlab.local -ZoneScope ntnxlab.local

Add the apps Ingress A record - use your own subdomain

Add-DnsServerResourceRecordA -Name *.apps.<ocpuserXX> -IPv4Address <your Ingress IP> -ZoneName ntnxlab.local -ZoneScope ntnxlab.local

Add the Prism Central A record

Add-DnsServerResourceRecordA -Name pc -IPv4Address <your PC IP> -ZoneName ntnxlab.local -ZoneScope ntnxlab.local

Sample commands with 'ocpuser01' as a subdomain and your OCP cluster name

Add-DnsServerResourceRecordA -Name api.ocpuser01 -IPv4Address 10.38.18.219 -ZoneName ntnxlab.local -ZoneScope ntnxlab.local
Add-DnsServerResourceRecordA -Name *.apps.ocpuser01 -IPv4Address 10.38.18.220 -ZoneName ntnxlab.local -ZoneScope ntnxlab.local

Add the Prism Central A record

Add-DnsServerResourceRecordA -Name pc -IPv4Address 10.38.18.201 -ZoneName ntnxlab.local -ZoneScope ntnxlab.local

Test name resolution for added entries

nslookup api.ocpuser01.ntnxlab.local
Server: dc.ntnxlab.local
Address: 10.38.18.203

Name: api.ocpuser01.ntnxlab.local
Address: 10.38.18.219 

nslookup myapp.apps.ocpuser01.ntnxlab.local
Server: dc.ntnxlab.local
Address: 10.38.18.203

Name: myapp.apps.ocpuser01.ntnxlab.local
Address: 10.38.18.220

nslookup pc.ntnxlab.local
Server: dc.ntnxlab.local
Address: 10.38.18.203

Name: pc.ntnxlab.local
Address: 10.38.3.201

Provision OCP Cluster in Red Hat Console

Go to Red Hat Console
Click on Create new assisted cluster
Fill in the following details:
- Cluster name - ocpuserXX (e.g. ocpuser01)
- Base domain - ntnxlab.local
- OpenShift version - 4.12.15
- CPU Architecture - x86_64
Click on Next
Click on Next in the Operators page
Click on Add Host
In the Add Host pop-up window select Minimal image file: Provision with virtual media
In the SSH public key text box provide the public key
Instructions to create SSH key pair
Logon to your UserXX-LinuxToolsVM
Username: root
Password: nutanix default
ssh-keygen -t rsa -b 2048 -f ~/.ssh/ocp # << accept all default answers ls -l ~/.ssh/ ## Output here ~/.ssh/ocp.pub ~/.ssh/ocp # << use this private key in the blueprints
Make sure to copy and paste the public key (ocp.pub)
Click on Generate Discovery ISO
Copy the Discovery ISO URL and note it down somewhere. You will need this for your next section while creating infrastructure.
Click on Close

Provision OCP Infrastructure

This section of the lab is done on using Linux Tools VM.

If it is not already present in your HPOC, create Linux Tools VM using instructions in create Linux Tools VM lab.

note

You are able to create these VMs and its resources using the Prism Element GUI. But in this section we will use Terraform code for repeatability and ease.

We will create the following resources:

OCP Role	Operating System	vCPU	RAM	Storage	IOPS
Master x 3	RHCOS	4	16 GB	100 GB	300
Worker x 2	RHCOS, RHEL 7.9, or RHEL 8.4	8	16 GB	100 GB	300

Run the following commands in sequence

yum update -y 
yum install -y yum-utils
yum-config-manager --add-repo https://rpm.releases.hashicorp.com/RHEL/hashicorp.repo
yum -y install terraform
yum -y install git

Clone the following git repo and initialise Terraform provider

git clone https://github.com/nutanix-japan/tf-ocp-infra
cd tf-ocp-infra
alias "tf=terraform" 
tf init

Get your variables file ready with your HPOC/SPOC environment information
```
cp terraform.tfvars.sample terraform.tfvars
```

Modify your variables to suit your HPOC environment using vi or nano editor

vi terraform.tfvars

Template file
Sample file

cluster_name        = "Prism Element Name" # << Change this
subnet_name         = "Primary"
user                = "admin"            
password            = "XXXXXXX"           # << Change this
endpoint            = "Prism Element IP"  # << Change this
vm_worker_prefix    = "ocpuserXX-worker"        # << Change xyz to your initials
vm_master_prefix    = "ocpuserXX-master"        # << Change xyz to your initials
vm_domain           = "ntnxlab.local"
vm_master_count     = 3
vm_worker_count     = 2
image_uri           = "Discover ISO URL you copied earlier" # << Change this

cluster_name        = "PHX-POC2xx"
subnet_name         = "Primary"
user                = "admin"
password            = "xxxxx"
endpoint            = "10.x.x.x"
vm_worker_prefix    = "ocpuser01-worker"
vm_master_prefix    = "ocpuser01-master"
vm_domain           = "ntnxlab.local"
vm_master_count     = 3
vm_worker_count     = 2
image_uri           = "https://api.openshift.com/api/assisted-images/images/78...."

Validate and run your Terraform code

tf validate

# if there are any validate errors fix or request support from your instructor

tf apply 

# Terraform will show you all resources that it will to create
# Type yes to confirm 

# Sample output for the command above

Terraform will perform the actions described above.
Only 'yes' will be accepted to approve.

Enter a value: yes

nutanix_image.RHCOS: Creating...
nutanix_image.RHCOS: Still creating... [10s elapsed]
nutanix_image.RHCOS: Creation complete after 14s [id=e04cff4e-a6cf-45f9-890d-96097c4b53ed]
nutanix_virtual_machine.RHCOS-master[0]: Creating...
nutanix_virtual_machine.RHCOS-master[1]: Creating...
nutanix_virtual_machine.RHCOS-worker[0]: Creating...
nutanix_virtual_machine.RHCOS-master[2]: Creating...
nutanix_virtual_machine.RHCOS-worker[1]: Creating...
nutanix_virtual_machine.RHCOS-master[0]: Still creating... [10s elapsed]
nutanix_virtual_machine.RHCOS-master[1]: Still creating... [10s elapsed]
nutanix_virtual_machine.RHCOS-master[2]: Still creating... [10s elapsed]
nutanix_virtual_machine.RHCOS-worker[0]: Still creating... [10s elapsed]
nutanix_virtual_machine.RHCOS-worker[1]: Still creating... [10s elapsed]
nutanix_virtual_machine.RHCOS-master[0]: Creation complete after 16s [id=3a88a3d7-304e-4284-886d-f7882764d7cc]
nutanix_virtual_machine.RHCOS-master[2]: Creation complete after 17s [id=5e87599a-5643-465d-9870-5b34751b2158]
nutanix_virtual_machine.RHCOS-worker[0]: Creation complete after 17s [id=78fb2e69-fee7-4244-ae5c-55ffbc1da21d]
nutanix_virtual_machine.RHCOS-master[1]: Creation complete after 17s [id=7775b527-fc55-4cac-aabc-a024ea4938c1]
nutanix_virtual_machine.RHCOS-worker[1]: Creation complete after 17s [id=c9801a82-a7e3-444e-a206-d5e3e3a75bb1]

Apply complete! Resources: 6 added, 0 changed, 0 destroyed.

Run the Terraform state list command to verify what resrouces are created

tf state list

# Sample output for the command above

data.nutanix_cluster.cluster            # < This is your existing Prism Element HPOC/SPOC cluster
data.nutanix_subnet.subnet              # < This is your existing Primary subnet
nutanix_image.RHCOS                     # < This is OCP Discovery ISO image
nutanix_virtual_machine.RHCOS-master[0] # < This is master vm 1
nutanix_virtual_machine.RHCOS-master[1] # < This is master vm 2
nutanix_virtual_machine.RHCOS-master[2] # < This is master vm 3
nutanix_virtual_machine.RHCOS-worker[0] # < This is worker vm 1
nutanix_virtual_machine.RHCOS-worker[1] # < This is worker vm 2

Install OCP Cluster in Red Hat Console

In this section we will use Red Hat Console's Assisted Installer wizard to install the OCP cluster with the VMs we have provisioned.

Return to Red Hat Openshift Console and check if the VMs appear (this may take up to 5 minutes)
In the Host inventory section, choose the following:
- Control Plane Node for Master VMs (the 3 Master VMs) and
- Worker nodes for Worker VMs from the drop-down menu (the 2 Worker VMs)
Click Next at the bottom of the page
In the Storage section, after confirming disk size, click Next at the bottom of the page
In the Networking section, choose Open Virtual Networking (OVN)
Assign IPs for your API Virtual IP and Ingress Virtual IP from you HPOC/SPOC Primary network pool that you reserved earlier in the AHV Networking section.
Click Next at the bottom of the page
Review your setup information and click on Install Cluster
You will be taken to monitoring your installation progress
Now the cluster deploy will proceed
Watch for any messages about user interactions in the progress page
This message is wanting the user to unmount the installation Discovery ISO so they VM can boot into the OS drive
Go to Prism Element > VM > Master/Worker VM > update
Under Disks > Click on Eject
Click on Save
Under Power Off Actions choose to Guest Reboot the VM where there are pending user action
Repeat ejecting CD-ROM for all VMs and rebooting it as the Wizard prompts for user action (do not do this before the prompting)
Once all the user actions are sustained for Master and Worker VMs, OCP cluster will be installed
Make sure to do the following:
- Download kubeconfig file
- Note down kubeadmin user password
- Note down the URL of the installed OCP cluster URL
note
This ocp cluster console URL can only be accessed within your HPOC/SPOC environment

Copy/scp the contents of kubeconfig file to your UserXX-LinuxToolsVM

Download and install oc command
curl -LO https://mirror.openshift.com/pub/openshift-v4/x86_64/clients/ocp/stable/openshift-client-linux.tar.gz
tar xvfz openshift-client-linux.tar.gz
cp oc kubectl /usr/local/bin/

export KUBECONFIG=/path/to/kubeconfig-file
oc get nodes
# You are all set

You have successfully deployed a OCP cluster using Red Hat Console Assisted Installer.

Takeaways

You can easily deploy multinode/single node OCP cluster using the Red Hat console
You can provision resoruces (VM, Storage, etc) on Nutanix using Terraform IaaC (GitOps)
info
The Installer Provisioned Installer (IPI) also uses Terraform to deploy infrastructure assets (VM) on Nutanix and VMware
Assisted Installer provisioned OCP clusters can be used as a learning ground and for testing purposes

note

Now that you have finished Assisted install, you can move on to the OCP Command Access section and then on to workloads section.

OCP: Assisted Installation

Pre-requisites for Assisted Installation​

High Level Overview of Assisted Installation​

OCP Assisted Installation​

Pre-requisites​

AHV Networking​

Add DNS Records​

Provision OCP Cluster in Red Hat Console​

Provision OCP Infrastructure​

Install OCP Cluster in Red Hat Console

Takeaways​

Pre-requisites for Assisted Installation

High Level Overview of Assisted Installation

OCP Assisted Installation

Pre-requisites

AHV Networking

Add DNS Records

Provision OCP Cluster in Red Hat Console

Provision OCP Infrastructure

Takeaways