Creating a ROSA cluster with PrivateLink enabled (custom VPC) and STS
This content is authored by Red Hat experts, but has not yet been tested on every supported configuration.
This is a combination of the private-link and sts setup documents to show the full picture

Prerequisites
AWS Preparation
- If this is a brand new AWS account that has never had a AWS Load Balancer installed in it, you should run the following - aws iam create-service-linked-role --aws-service-name \ "elasticloadbalancing.amazonaws.com"
Create the AWS Virtual Private Cloud (VPC) and Subnets
For this scenario, we will be using a newly created VPC with both public and private subnets. All of the cluster resources will reside in the private subnet. The public subnet will be used for traffic to the Internet (egress)
Note: If you already have a Transit Gateway (TGW) or similar, you can skip the public subnet configuration
Note: When creating subnets, make sure that subnet(s) are created in availability zones that have ROSA instances types available. If AZ is not “forced”, the subnet is created in a random AZ in the region. Force AZ using the
--availability-zoneargument in thecreate-subnetcommand.
- Use - rosa list instance-typesto list the ROSA instance types
- Use - aws ec2 describe-instance-type-offeringsto check that your desired AZ supports your desired instance type- Example using us-east-1, us-east-1b, and m5.xlarge: - aws ec2 describe-instance-type-offerings --location-type availability-zone \ --filters Name=location,Values=us-east-1b --region us-east-1 \ --output text | egrep m5.xlarge- Result should display INSTANCETYPEOFFERINGS [instance-type] [az] availability-zone if your selected region supports your desired instance type 
- Configure the following environment variables, adjusting for - ROSA_CLUSTER_NAME,- VERSIONand- REGIONas necessary- export VERSION=4.9.15 \ ROSA_CLUSTER_NAME=pl-sts-cluster \ AWS_ACCOUNT_ID=`aws sts get-caller-identity --query Account --output text` \ REGION=us-east-1 \ AWS_PAGER=""
- Create a VPC for use by ROSA - Create the VPC and return the ID as - VPC_ID- VPC_ID=`aws ec2 create-vpc --cidr-block 10.0.0.0/16 | jq -r .Vpc.VpcId` echo $VPC_ID
- Tag the newly created VPC with the cluster name - aws ec2 create-tags --resources $VPC_ID \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME
- Configure the VPC to allow DNS hostnames for their public IP addresses - aws ec2 modify-vpc-attribute --vpc-id $VPC_ID --enable-dns-hostnames
- The new VPC should be visible in the AWS console  
 
- Create a Public Subnet to allow egress traffic to the Internet - Create the public subnet in the VPC CIDR block range and return the ID as - PUBLIC_SUBNET- PUBLIC_SUBNET=`aws ec2 create-subnet --vpc-id $VPC_ID --cidr-block 10.0.128.0/17 | jq -r .Subnet.SubnetId` echo $PUBLIC_SUBNET
- Tag the public subnet with the cluster name - aws ec2 create-tags --resources $PUBLIC_SUBNET \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME-public
 
- Create a Private Subnet for the cluster - Create the private subnet in the VPC CIDR block range and return the ID as - PRIVATE_SUBNET- PRIVATE_SUBNET=`aws ec2 create-subnet --vpc-id $VPC_ID \ --cidr-block 10.0.0.0/17 | jq -r .Subnet.SubnetId` echo $PRIVATE_SUBNET
- Tag the private subnet with the cluster name - aws ec2 create-tags --resources $PRIVATE_SUBNET \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME-private
- Both subnets should now be visible in the AWS console  
 
- Create an Internet Gateway for NAT egress traffic - Create the Internet Gateway and return the ID as - I_GW- I_GW=`aws ec2 create-internet-gateway | jq -r .InternetGateway.InternetGatewayId` echo $I_GW
- Attach the new Internet Gateway to the VPC - aws ec2 attach-internet-gateway --vpc-id $VPC_ID --internet-gateway-id $I_GW
- Tag the Internet Gateway with the cluster name - aws ec2 create-tags --resources $I_GW \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME
- The new Internet Gateway should be created and attached to your VPC  
 
- Create a Route Table for NAT egress traffic - Create the Route Table and return the ID as - R_TABLE- R_TABLE=`aws ec2 create-route-table --vpc-id $VPC_ID \ | jq -r .RouteTable.RouteTableId` echo $R_TABLE
- Create a route with no IP limitations (0.0.0.0/0) to the Internet Gateway - aws ec2 create-route --route-table-id $R_TABLE \ --destination-cidr-block 0.0.0.0/0 --gateway-id $I_GW
- Verify the route table settings - aws ec2 describe-route-tables --route-table-id $R_TABLE- Example output  
- Associate the Route Table with the Public subnet - aws ec2 associate-route-table --subnet-id $PUBLIC_SUBNET \ --route-table-id $R_TABLE- Example output  
- Tag the Route Table with the cluster name - aws ec2 create-tags --resources $R_TABLE \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME
 
- Create a NAT Gateway for the Private network - Allocate and elastic IP address and return the ID as - EIP- EIP=`aws ec2 allocate-address --domain vpc | jq -r .AllocationId` echo $EIP
- Create a new NAT Gateway in the Public subnet with the new Elastic IP address and return the ID as - NAT_GW- NAT_GW=`aws ec2 create-nat-gateway --subnet-id $PUBLIC_SUBNET \ --allocation-id $EIP | jq -r .NatGateway.NatGatewayId` echo $NAT_GW
- Tag the Elastic IP with the cluster name - aws ec2 create-tags --resources $EIP --resources $NAT_GW \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME
- The new NAT Gateway should be created and associated with your VPC  
 
- Create a Route Table for the Private subnet to the NAT Gateway - Create a Route Table in the VPC and return the ID as - R_TABLE_NAT- R_TABLE_NAT=`aws ec2 create-route-table --vpc-id $VPC_ID \ | jq -r .RouteTable.RouteTableId` echo $R_TABLE_NAT
- Loop through a Route Table check until it is created - while ! aws ec2 describe-route-tables \ --route-table-id $R_TABLE_NAT \ | jq .; do sleep 1; done- Example output!  
- Create a route in the new Route Table for all addresses to the NAT Gateway - aws ec2 create-route --route-table-id $R_TABLE_NAT \ --destination-cidr-block 0.0.0.0/0 \ --gateway-id $NAT_GW
- Associate the Route Table with the Private subnet - aws ec2 associate-route-table --subnet-id $PRIVATE_SUBNET \ --route-table-id $R_TABLE_NAT
- Tag the Route Table with the cluster name - aws ec2 create-tags --resources $R_TABLE_NAT $EIP \ --tags Key=Name,Value=$ROSA_CLUSTER_NAME-private
 
Configure the AWS Security Token Service (STS) for use with ROSA
The AWS Security Token Service (STS) allows us to deploy ROSA without needing a ROSA admin account, instead it uses roles and policies to gain access to the AWS resources needed to install and operate the cluster.
This is a summary of the official OpenShift docs that can be used as a line by line install guide.
Note that some commands (OIDC for STS) will be hard coded to US-EAST-1, do not be tempted to change these to use $region instead or you will fail installation.
- Make you your ROSA CLI version is correct (v1.1.0 or higher) - rosa version
- Create the IAM Account Roles - rosa create account-roles --mode auto --yes
Deploy ROSA cluster
- Run the rosa cli to create your cluster - rosa create cluster -y --cluster-name ${ROSA_CLUSTER_NAME} \ --region ${REGION} --version ${VERSION} \ --subnet-ids=$PRIVATE_SUBNET \ --private-link --machine-cidr=10.0.0.0/16 \ --sts- Confirm the PrivateLink set up  
- Create the Operator Roles - rosa create operator-roles -c $ROSA_CLUSTER_NAME --mode auto --yes
- Create the OIDC provider. - rosa create oidc-provider -c $ROSA_CLUSTER_NAME --mode auto --yes
- Validate The cluster is now installing - The State should have moved beyond - pendingand show- installingor- ready.- watch "rosa describe cluster -c $ROSA_CLUSTER_NAME"
- Watch the install logs - rosa logs install -c $ROSA_CLUSTER_NAME --watch --tail 10
Validate the cluster
Once the cluster has finished installing it is time to validate. Validation when using PrivateLink requires the use of a jump host.
You can create them using the AWS Console or the AWS CLI as depicted below:
- Option 1: Create a jump host instance through the AWS Console - Navigate to the EC2 console and launch a new instance 
- Select the AMI for your instance, if you don’t have a standard, the Amazon Linux 2 AMI works just fine  
- Choose your instance type, the t2.micro/free tier is sufficient for our needs, and click Next: Configure Instance Details 
- Change the Network settings to setup this host inside your private-link VPC  
- Change the Subnet setting to use the private-link-public subnet  
- Change Auto-assign Public IP to Enable  
- Default settings for Storage and Tags are fine. Make the following changes in the 6. Configure Security Group tab (either by clicking through the screens or selecting from the top bar) - If you already have a security group created to allow access from your computer to AWS, choose Select an existing security group and choose that group from the list, otherwise, select Create a new security group and continue. 
- To allow access only from your current public IP, change the Source heading to use My IP  
 
- Click Review and Launch, verify all settings are correct, and follow the standard AWS instructions for finalizing the setup and selecting/creating the security keys. 
- Once launched, open the instance summary for the jump host instance and note the public IP address. 
 
- Option 2: Create a jumphost instance using the AWS CLI - Create an additional Security Group for the jumphost - TAG_SG="$ROSA_CLUSTER_NAME-jumphost-sg" aws ec2 create-security-group --group-name ${ROSA_CLUSTER_NAME}-jumphost-sg --description ${ROSA_CLUSTER_NAME}-jumphost-sg --vpc-id ${VPC_ID} --tag-specifications "ResourceType=security-group,Tags=[{Key=Name,Value=$TAG_SG}]"
- Grab the Security Group Id generated in the previous step - PublicSecurityGroupId=$(aws ec2 describe-security-groups --filters "Name=tag:Name,Values=${ROSA_CLUSTER_NAME}-jumphost-sg" | jq -r '.SecurityGroups[0].GroupId') echo $PublicSecurityGroupId
- Add a rule to Allow the ssh into the Public Security Group - aws ec2 authorize-security-group-ingress --group-id $PublicSecurityGroupId --protocol tcp --port 22 --cidr 0.0.0.0/0
- (Optional) Create a Key Pair for your jumphost if your have not a previous one - aws ec2 create-key-pair --key-name $ROSA_CLUSTER_NAME-key --query 'KeyMaterial' --output text > PATH/TO/YOUR_KEY.pem chmod 400 PATH/TO/YOUR_KEY.pem
- Define an AMI_ID to be used for your jump host - AMI_ID="ami-0022f774911c1d690"- This AMI_ID corresponds an Amazon Linux within the us-east-1 region and could be not available in your region. Find your AMI ID and use the proper ID. 
- Launch an ec2 instance for your jumphost using the parameters defined in early steps: - TAG_VM="$ROSA_CLUSTER_NAME-jumphost-vm" aws ec2 run-instances --image-id $AMI_ID --count 1 --instance-type t2.micro --key-name $ROSA_CLUSTER_NAME-key --security-group-ids $PublicSecurityGroupId --subnet-id $PUBLIC_SUBNET --associate-public-ip-address --tag-specifications "ResourceType=instance,Tags=[{Key=Name,Value=$TAG_VM}]"- This instance will be associated with a Public IP directly. 
- Wait until the ec2 instance is in Running state, grab the Public IP associated to the instance and check the if the ssh port and: - IpPublicBastion=$(aws ec2 describe-instances --filters "Name=tag:Name,Values=$TAG_VM" | jq -r '.Reservations[0].Instances[0].PublicIpAddress') echo $IpPublicBastion nc -vz $IpPublicBastion 22
 
- Create a ROSA admin user and save the login command for use later - rosa create admin -c $ROSA_CLUSTER_NAME
- Note the DNS name of your private cluster, use the - rosa describecommand if needed- rosa describe cluster -c $ROSA_CLUSTER_NAME
- update /etc/hosts to point the openshift domains to localhost. Use the DNS of your openshift cluster as described in the previous step in place of - $YOUR_OPENSHIFT_DNSbelow- 127.0.0.1 api.$YOUR_OPENSHIFT_DNS 127.0.0.1 console-openshift-console.apps.$YOUR_OPENSHIFT_DNS 127.0.0.1 oauth-openshift.apps.$YOUR_OPENSHIFT_DNS
- SSH to that instance, tunneling traffic for the appropriate hostnames. Be sure to use your new/existing private key, the OpenShift DNS for - $YOUR_OPENSHIFT_DNSand your jump host IP for- $YOUR_EC2_IP- sudo ssh -i PATH/TO/YOUR_KEY.pem \ -L 6443:api.$YOUR_OPENSHIFT_DNS:6443 \ -L 443:console-openshift-console.apps.$YOUR_OPENSHIFT_DNS:443 \ -L 80:console-openshift-console.apps.$YOUR_OPENSHIFT_DNS:80 \ ec2-user@$YOUR_EC2_IP 
- From your EC2 jump instances, download the OC CLI and install it locally - Download the OC CLI for Linuxwget https://mirror.openshift.com/pub/openshift-v4/clients/ocp/stable/openshift-client-linux.tar.gz
- Unzip and untar the binarygunzip openshift-client-linux.tar.gz tar -xvf openshift-client-linux.tar
 
- Download the OC CLI for Linux
- log into the cluster using oc login command from the create admin command above. ex. - ./oc login https://api.$YOUR_OPENSHIFT_DNS.p1.openshiftapps.com:6443 --username cluster-admin --password $YOUR_OPENSHIFT_PWD 
- Check that you can access the Console by opening the console url in your browser.  
Cleanup
- Delete ROSA - rosa delete cluster -c $ROSA_CLUSTER_NAME -y
- Watch the logs and wait until the cluster is deleted - rosa logs uninstall -c $ROSA_CLUSTER_NAME --watch
- Clean up the STS roles - Note you can get the correct commands with the ID filled in from the output of the previous step. - rosa delete operator-roles -c <id> --mode auto --yes rosa delete oidc-provider -c <id> --mode auto --yes
- Delete AWS resources - aws ec2 delete-nat-gateway --nat-gateway-id $NAT_GW | jq . aws ec2 release-address --allocation-id=$EIP | jq . aws ec2 detach-internet-gateway --vpc-id $VPC_ID --internet-gateway-id $I_GW | jq . aws ec2 delete-subnet --subnet-id=$PRIVATE_SUBNET | jq . aws ec2 delete-subnet --subnet-id=$PUBLIC_SUBNET | jq . aws ec2 delete-route-table --route-table-id=$R_TABLE | jq . aws ec2 delete-route-table --route-table-id=$R_TABLE_NAT | jq . aws ec2 delete-internet-gateway --internet-gateway-id $I_GW | jq . aws ec2 delete-vpc --vpc-id=$VPC_ID | jq .