Skip to content

Latest commit

 

History

History
172 lines (128 loc) · 5.22 KB

module-2-deploy-aks.md

File metadata and controls

172 lines (128 loc) · 5.22 KB
Raw

Module 2 - Deploy an Azure AKS Cluster

Initial Setup

  1. Define the environment variables to be used by the resources definition.

    NOTE: In the following sections we'll be generating and setting some environment variables. If you're terminal session restarts you may need to reset these variables. You can use that via the following command:

    source envLabVars.env

    Start in the root of the repository folder

    cd cc-aks-ebpf
    export RESOURCE_GROUP=rg-cc-aks-ebpf
export CLUSTERNAME=myname-cc-aks-ebpf
export LOCATION=canadacentral
export K8S_VERSION=1.30.6
export POD_CIDR='10.244.0.0/16'
# Persist for later sessions in case of disconnection.
echo export RESOURCE_GROUP=$RESOURCE_GROUP >> envLabVars.env
echo export CLUSTERNAME=$CLUSTERNAME >> envLabVars.env
echo export LOCATION=$LOCATION >> envLabVars.env
echo export K8S_VERSION=$K8S_VERSION >> envLabVars.env
echo export POD_CIDR=$POD_CIDR >> envLabVars.env

  2. If not created, create the Resource Group in the desired Region.

    az group create \
  --name $RESOURCE_GROUP \
  --location $LOCATION

  3. Create the AKS cluster without a network plugin. We will install Calico OSS CNI afterwards.

    az aks create \
  --resource-group $RESOURCE_GROUP \
  --name $CLUSTERNAME \
  --kubernetes-version $K8S_VERSION \
  --nodepool-name 'nodepool1' \
  --location $LOCATION \
  --node-count 3 \
  --network-plugin none \
  --pod-cidr $POD_CIDR \
  --node-osdisk-size 50 \
  --node-vm-size Standard_B2ms \
  --max-pods 70 \
  --generate-ssh-keys \
  --enable-managed-identity \
  --output table

  4. Verify your cluster status. The ProvisioningState should be Succeeded

    az aks list -o table | grep $CLUSTERNAME

    Or

    watch az aks list -o table

    You may get an output like the following

    Name                     Location       ResourceGroup              KubernetesVersion    CurrentKubernetesVersion    ProvisioningState    Fqdn
-----------------------  -------------  -------------------------  -------------------  --------------------------  -------------------  -----------------------------------------------------------------------
aks-cc-aks-ebpf                canadacentral  rg-cc-aks-ebpf                1.30                 1.30.6                      Succeeded          aks-rg-cc-aks-03cfb8-ub5gqil0.hcp.canadacentral.azmk8s.io

  5. Get the credentials to connect to the cluster.

    az aks get-credentials --resource-group $RESOURCE_GROUP --name $CLUSTERNAME

  6. Verify you have API access to your new AKS cluster

    kubectl get nodes

    The output will be something similar to the this:

     NAME                                STATUS   ROLES   AGE    VERSION
 aks-nodepool1-24664026-vmss000000   NotReady    agent   7m1s   v1.30.6
 aks-nodepool1-24664026-vmss000001   NotReady    agent   7m4s   v1.30.6

    The nodes are showing NotReady because we haven't actually installed Calico as a CNI yet, which is the next step.

Install Calico CNI in iptables/nftables mode

By default, when installing Calico as a CNI on an AKS cluster as a BYOCNI option, the default dataplane will be iptables/nftables (depending on the OS) unless the user configures it as eBPF. For the sake of this workshop, we want to compare the two dataplanes so we start with iptables and then convert the cluster to eBPF later.

  1. Configure Calico Helm repo

    helm repo add projectcalico https://docs.tigera.io/calico/charts

  2. Configure Helm values

    cat > values.yaml <<EOF
installation:
  kubernetesProvider: AKS
  cni:
   type: Calico
   ipam:
     type: Calico
  calicoNetwork:
    hostPorts: Disabled
    bgp: Disabled
    ipPools:
    - cidr: $POD_CIDR
      encapsulation: VXLAN
EOF

  3. Install Calico OSS using Helm values file

    kubectl create namespace tigera-operator
    helm install calico projectcalico/tigera-operator --version v3.29.1 -f values.yaml --namespace tigera-operator

  4. The nodes should now be networked properly with Calico as CNI and should show status as Ready

     NAME                                STATUS   ROLES   AGE    VERSION
 aks-nodepool1-24664026-vmss000000   Ready    agent   9m3s   v1.30.6
 aks-nodepool1-24664026-vmss000001   Ready    agent   9m3s   v1.30.6

  5. At this point, the Calico components should also be functional and show as Available

    kubectl get tigerastatus

    The output should look something like this:

    NAME                            AVAILABLE   PROGRESSING   DEGRADED   SINCE
apiserver                       True        False         False      11m
calico                          True        False         False      11m

Now we are ready to connect the cluster to Calico Cloud.

➡️ Module 3 - Connect the cluster to Calico Cloud

⬅️ Module 1 - Getting Started

↩️ Back to Main