An administrator is planning to deploy a new Azure Kubernetes Service (AKS) cluster named `aks-finance`. The cluster must integrate with an on-premises network, and pods must be directly reachable from the on-premises network using their own IP addresses from the Azure Virtual Network (VNet) without any network address translation (NAT). The network security team has allocated a small subnet with a prefix for the cluster nodes, but wants to ensure that pod deployments do not quickly exhaust the IP addresses in this node subnet.
Which two configuration options should the administrator select to meet these requirements? (Select two.)
- Configure the cluster to use the Azure CNI network plugin.Answer
- BConfigure the cluster to use the Kubenet network plugin.
- Enable dynamic pod IP allocation to allocate pod IP addresses from a separate subnet.Answer
- DConfigure the cluster to use Azure CNI Overlay.
- EDeploy the cluster nodes and pods in the same subnet with a standard Kubenet configuration and configure User-Defined Routes (UDRs) for on-premises routing.
Answer
Configure the cluster to use the Azure CNI network plugin and enable dynamic pod IP allocation to allocate pod IP addresses from a separate subnet.
To ensure pods have direct connectivity to the on-premises network without NAT, the Azure CNI network plugin must be used since it assigns VNet IPs directly to pods. To prevent the node subnet from being exhausted by pod IP allocations, dynamic pod IP allocation should be enabled. This allows nodes and pods to reside in different subnets, ensuring that the node subnet is not exhausted by the pod density.
Step-by-Step Solution
Key Concept
Azure CNI vs Kubenet, and Dynamic Pod IP Allocation in Azure Kubernetes Service (AKS)