Prepare and test your skills

Prepare and test your skills

Free sample question with answers and detailed explanation below. The Professional Cloud Network Engineer practice bank has an additional 2,437 practice questions.Try 10 free realistic questions: instant score, no credit card, and no account required.
You are designing a VPC-native GKE Standard cluster in a Shared VPC network with strictly constrained RFC 1918 IP address space. A network administrator will pre-create the subnet and user-managed secondary ranges to meet these maximum scale requirements:
kube-system Pods)ClusterIP ServicesWhich CIDR sizes and cluster configuration allocate the smallest valid IP ranges while supporting full cluster scale?
By default, GKE Standard clusters configure --default-max-pods-per-node=110. To guarantee at least twice as many IP addresses as running Pods per node (2 × 110 = 220, rounded up to the next power of two), GKE carves out a /24 Pod CIDR block (256 IP addresses) for every node that joins the cluster.
/19 secondary Pod range contains 8,192 IP addresses (2^(32 - 19)). When each node reserves a /24 (256 IPs), the /19 range supports at most 8,192 / 256 = 32 nodes./24 block. The cluster reaches 100% Pod range utilization at 32 nodes, blocking autoscaling well short of the required 120 nodes.max-pods-per-node value cannot be modified on an existing cluster or node pool after creation.--default-max-pods-per-node=110 would require 120 × 256 = 30,720 Pod IPs—necessitating a /17 secondary Pod range (32,768 IPs) instead of a /19.max-pods-per-node wastes up to 75% of the Pod IP space when actual Pod density is 32 or fewer Pods per node.--max-pods-per-node, cordoning and draining workloads, and deleting the default pool.The cluster only needed to scale up to 32 nodes while running higher-density workloads (up to 110 Pods per node), or if a /17 Pod secondary range was allocated instead of a /19 without strict IP conservation constraints.
A /25 primary IPv4 subnet contains 128 total IP addresses (2^(32 - 25)). In VPC networks, this range supplies the internal IP addresses for GKE worker nodes as well as any internal TCP/UDP LoadBalancer Services created without a dedicated load balancer subnet.
/25 therefore provides only 128 - 4 = 124 usable IP addresses.LoadBalancer Services allocate VIPs from the node subnet's primary range. Supporting 120 nodes plus 8 internal load balancers requires 128 usable primary IP addresses, exceeding the 124 usable addresses in a /25.expand-ip-range), but only if adjacent contiguous IP space is unallocated in the Shared VPC.--private-endpoint-subnetwork, 1 additional primary IP is consumed by the control plane endpoint, leaving only 123 node IPs in a /25.The 8 internal TCP/UDP LoadBalancer Services were placed in a separate dedicated load balancer subnet (or fewer than 5 internal load balancers were deployed), leaving all 120 nodes within the 124 usable IPs of a /25.
A /20 secondary IPv4 range provides 4,096 usable Pod IP addresses (2^(32 - 20)). Unlike primary subnet ranges, all IP addresses in a secondary range are available for Pod allocation without subtracting 4 reserved addresses.
--default-max-pods-per-node (2^(ceil(log2(2 × Q)))) to prevent IP reuse collisions during rapid Pod creation and termination.--default-max-pods-per-node=32, each node consumes a /26 block (64 IP addresses). Dividing a /20 (4,096 IPs) by 64 IPs per node yields a hard ceiling of only 64 nodes (4,096 / 64 = 64).120 × 32 = 3,840 individual Pods is numerically less than 4,096, the cluster will hit 100% Pod IP allocation at 64 nodes and throw IP_SPACE_EXHAUSTED errors before scaling to 120 nodes./26 block as allocated regardless of how many Pods are running on that node.--max-pods-per-node.The cluster only needed to autoscale to 64 nodes (rather than 120 nodes), or if workloads required at most 16 Pods per node (--default-max-pods-per-node=16), which assigns a /27 (32 IPs) per node and allows 4,096 / 32 = 128 nodes in a /20.
This configuration combines a /24 primary subnet range, a /19 secondary Pod range paired with --default-max-pods-per-node=32, and a /23 secondary Service range to provision exact, non-exhaustible alias IP capacity for a Shared VPC-native GKE Standard cluster.
/24 primary range): Google Cloud reserves 4 IP addresses (the first two and last two) in every primary subnet range (N = 2^(32 - S) - 4). Because internal TCP/UDP LoadBalancer Services consume IP addresses from the node subnet by default, the cluster requires 120 + 8 = 128 usable primary IPs. A /25 yields only 128 - 4 = 124 usable IPs, whereas a /24 yields 256 - 4 = 252 usable IPs./19 Pod range with --default-max-pods-per-node=32): GKE assigns each node an exclusive Pod CIDR block sized to the smallest power of two holding at least 2 × max-pods-per-node (M = 31 - ceil(log2(Q))). Setting Q = 32 allocates a /26 block (64 Pod IPs) per node. Supporting 120 nodes requires 120 × 64 = 7,680 Pod IPs. All IPs in a secondary range are usable, so a /19 (8,192 IPs) supports up to 128 nodes (8,192 / 64)./23 Service range): A /24 provides only 256 ClusterIP addresses, while a /23 provides 512 addresses, accommodating 500 Services.IP_SPACE_EXHAUSTED failures when the cluster autoscaler scales up to 120 nodes./26) to actual workload density.Any smaller primary (/25), Pod (/20), or Service (/24) prefix fails to meet the stated scale limits, while leaving --default-max-pods-per-node at 110 would require a /17 Pod range (32,768 IPs) for 120 nodes—wasting over 24,000 IP addresses.
Answer them yourself and get your score the moment you finish. No account needed.
Full access includes
Keep the momentum going with these hand-picked practice scenarios
Last updated Validated