With Public Endpoints

Use this configuration when you want the API endpoints of your OKE clusters to be public.

• Create the clusters
• Set up kubeconfig
• Install Verrrazzano

Create the clusters

1. Copy terraform.tfvars.example to terraform.tfvars:

cp terraform.tfvars.example terraform.tfvars

Provider

1. Provide values for the following input variables in terraform.tfvars:

   • api_fingerprint
   • api_private_key_path
   • tenancy_id
   • user_id
   • compartment_id
   • ssh_public_key_path
   • ssh_private_key_path

2. Set your tenancy's home region using home_region in terraform.tfvars e.g.

home_region = "us-ashburn-1"

Admin Cluster

1. Set the Admin region for where you want to place the Verrazzano Admin cluster. This should be the region name only e.g. Sydney. This should not be the region identifier ap-sydney-1 e.g.

   admin_region = {
     admin_name = "admin"
     region     = "Sydney" # must match the provider's region in step 4
     vcn_cidr   = "10.0.0.0/16"
     pods       = "10.200.0.0/16"
     services   = "10.100.0.0/16"
   }
   

2. Change the provider of the Admin cluster to its corresponding alias. The aliases can be found in providers.tf e.g. if you want to place your Admin cluster in Ashburn, you must change the provider from oci.sydney to oci.ashburn. This should be configured in admin.tf in the root module e.g.

  providers = {
    oci      = oci.ashburn
    oci.home = oci.home
  }

Managed Clusters

1. To create managed clusters in your preferred regions, set your preferred regions' values to true e.g. ashburn=true in the clusters parameter. Leave those that you do not use to false e.g.

clusters = {
  # Africa
  johannesburg = false

  # Asia
  chuncheon = false
  hyderabad = false
  mumbai    = false
  osaka     = false
  seoul     = false
  singapore = false
  tokyo     = false

  # Europe
  amsterdam = false
  frankfurt = false
  london    = false
  madrid    = false
  marseille = false
  milan     = false
  newport   = false
  paris     = false
  stockholm = false
  zurich    = false

  # Middle East
  abudhabi  = false
  dubai     = false
  jeddah    = false
  jerusalem = false

  # Oceania
  melbourne = false
  sydney    = false

  # South America
  santiago = false
  saupaulo = false
  vinhedo  = false

  # North America
  ashburn   = false
  chicago   = false
  montreal  = false
  phoenix   = true
  queretaro = false
  sanjose   = false
  toronto   = false
}

2. For the managed clusters you enable, uncomment them in modules\clusters\outputs.tf for cluster_ids, int_nsg_ids and pub_nsg_ids. For the clusters that you do not use, leave them commented e.g.

output "cluster_ids" {
  value = {
    # "johannesburg" = one(module.johannesburg[*].cluster_id)
    # "chuncheon"    = one(module.chuncheon[*].cluster_id)
    # "hyderabad"    = one(module.hyderabad[*].cluster_id)
    # "mumbai"       = one(module.mumbai[*].cluster_id)
    # "osaka"        = one(module.osaka[*].cluster_id)
    # "seoul"        = one(module.seoul[*].cluster_id)
    # "singapore"    = one(module.singapore[*].cluster_id)
    # "tokyo"        = one(module.tokyo[*].cluster_id)
    # "amsterdam"    = one(module.amsterdam[*].cluster_id)
    # "frankfurt"    = one(module.frankfurt[*].cluster_id)
    # "london"       = one(module.london[*].cluster_id)
    # "madrid"       = one(module.madrid[*].cluster_id)
    # "marseille"    = one(module.marseille[*].cluster_id)
    # "milan"        = one(module.milan[*].cluster_id)
    # "newport"      = one(module.newport[*].cluster_id)
    # "paris"        = one(module.paris[*].cluster_id)
    # "stockholm"    = one(module.stockholm[*].cluster_id)
    # "zurich"       = one(module.zurich[*].cluster_id)
    # "abudhabi"     = one(module.abudhabi[*].cluster_id)
    # "dubai"        = one(module.dubai[*].cluster_id)
    # "jeddah"       = one(module.jeddah[*].cluster_id)
    # "jerusalem"    = one(module.jerusalem[*].cluster_id)
    # "melbourne"    = one(module.melbourne[*].cluster_id)
    # "sydney"       = one(module.sydney[*].cluster_id)
    "melbourne" = one(module.melbourne[*].cluster_id)
    # "sydney"    = one(module.sydney[*].cluster_id)
    # "santiago"     = one(module.santiago[*].cluster_id)
    # "saupaulo"     = one(module.saupaulo[*].cluster_id)
    # "vinhedo"      = one(module.vinhedo[*].cluster_id)
    # "ashburn"      = one(module.ashburn[*].cluster_id)
    # "chicago"      = one(module.chicago[*].cluster_id)
    # "montreal"     = one(module.montreal[*].cluster_id)
    # "phoenix"      = one(module.phoenix[*].cluster_id)
    # "queretaro"    = one(module.queretaro[*].cluster_id)
    # "sanjose"      = one(module.sanjose[*].cluster_id)
    # "toronto"      = one(module.toronto[*].cluster_id)
  }
}

3. Uncomment the respective clusters that you have enabled. The clusters are created in either {country, continent,region}.tf. e.g. Ashburn will be found in usa.tf under modules\clusters. Leave the clusters that you do not use as commented.

4. Configure the API endpoints for public access in terraform.tfvars:

oke_control_plane = "public"

5. When initially creating the clusters, configure the following input variables to be false:

get_kubeconfigs = false
install_verrazzano = false

6. Create the clusters by running Terraform:

terraform init
terraform plan
terraform apply

7. After Terraform has finished the apply operation, it will output a convenient SSH command you can copy. This command will allow you to ssh to the operator host.

Set up kubeconfig

You will now generate the installation script.

1. Configure the following input variables:

get_kubeconfigs = true
install_verrazzano = true

2. Configure the Verrazzano profile:

verrazzano_profile = "prod"

3. If you need to use your own DNS, you need to configure DNS before you run Terraform and the installation scripts. Else, the installation scripts will use nip.io.

4. If you want to configure additional Verrazzano components, you can still do so now.

5. Run Terraform again:

terraform init
terraform plan
terraform apply

5. ssh to the operator host using the convenience command printed by Terraform.

ssh -i /path/to/ssh_private_key -J opc@xyz.xyz.xyz.xyz opc@abc.abc.abc.abc"

6. Verify you can reach the clusters e.g.

for cluster in admin phoenix ; do
  kubectx $cluster
  kubectl get nodes
done

7. You should be able to see something like the following:

    ✔ Switched to context "admin".
    NAME          STATUS   ROLES   AGE   VERSION
    10.0.123.29   Ready    node    91m   v1.24.1
    ✔ Switched to context "phoenix".
    NAME            STATUS   ROLES   AGE   VERSION
    10.31.101.163   Ready    node    89m   v1.24.1
    10.31.84.171    Ready    node    89m   v1.24.1

Install Verrrazzano

You will now install Verrazzano in all clusters and register the managed clusters with the Admin cluster.

1. Install the Verrazzano Platform Operator:

cd /home/opc/vz/operator

for cluster in admin phoenix ; do
  bash install_vz_operator_$cluster.sh
done

2. Check if the operator has installed successfully in all clusters:

for cluster in admin phoenix; do
  bash check_vz_operator_$cluster.sh
done

3. Obtain the manifest for DNS. Follow the instructions to configure a secret for DNS.

cd /home/opc/vz/oci
bash  get_oci_secret.sh
echo >> oci.yaml

4. Create the secret in each cluster:

for cluster in admin phoenix ; do
  bash create_oci_secret_$cluster.sh
done

5. If you are using Thanos, configure your secret for OCI Object Storage now. For each enabled cluster, a storage file is created. Edit them and enter the following and create a secret:

• compartment_ocid
• region
• tenancy_ocid
• user_ocid
• fingerprint
• and the private key

cd /home/opc/vz/clusters
for cluster in admin phoenix; do
  kubectx $cluster
  kubectl create namespace verrazzano-monitoring
  kubectl create secret generic objstore-config -n verrazzano-monitoring --from-file=objstore.yml=thanos_${cluster}_storage.yaml
done

6. Install the Verrazzano Admin cluster:

cd /home/opc/vz/clusters
bash install_vz_cluster_admin.sh

7. While the Admin cluster is being installed in the background, you can install the managed clusters in parallel:

cd /home/opc/vz/clusters
for cluster in phoenix ; do
  bash install_vz_cluster_$cluster.sh
done

8. Wait for Verrazzano to be installed in all clusters:

# check managed clusters' status
bash vz_status.sh

# Admin cluster takes longer, check separately
kubectx admin
kubectl wait --timeout=20m --for=condition=InstallComplete verrazzano/admin

9. Create the certificates secrets for each managed cluster:

cd /home/opc/vz/certs
for cluster in phoenix; do
  bash create_cert_secret_$cluster.sh
done

10. Create the ConfigMap for the API Server:

cd /home/opc/vz/cm
bash create_api_cm.sh

11. Create the Verrazzano managed cluster objects for each managed cluster:

cd /home/opc/vz/clusters
for cluster in phoenix; do
  bash create_vmc_$cluster.sh
done

12. Register all the managed clusters:

for cluster in phoenix; do
  bash register_vmc_$cluster.sh
done