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.
Your organization manages a production VPC-native GKE Standard cluster in a Shared VPC with control plane authorized networks enabled. The network topology currently uses the following CIDR allocations:
| Resource | CIDR Range |
|---|---|
| Primary node subnet | 10.20.0.0/24 |
| Pod secondary range | 10.24.0.0/16 |
| Service secondary range | 10.28.0.0/20 |
| Control plane range | 172.16.0.0/28 |
| Current on-premises admin range (authorized) | 192.168.10.0/24 |
| On-premises branch range (via Cloud Interconnect) | 10.20.2.0/24 |
You need to scale the cluster to support 450 nodes and transition administrators to a new on-premises CIDR (192.168.20.0/24) before decommissioning 192.168.10.0/24. You must avoid IP overlaps with existing routes and prevent any disruption to running workloads, node-to-control-plane communication, or active administrator access.
Which sequence of actions should you implement?
In a VPC-native Google Kubernetes Engine (GKE) cluster, in-place primary subnet expansion increases the IPv4 address space available for cluster nodes without recreating the subnet or disrupting active workloads. When control plane authorized networks are enabled, the list of permitted CIDR blocks acts as an allowlist at the Kubernetes API server boundary and must explicitly encompass the expanded node subnet as well as all active administrative networks.
10.20.0.0/24 to 10.20.0.0/23 (10.20.0.0 through 10.20.1.255) yields 512 total addresses (508 usable after subtracting Google Cloud's 4 reserved addresses). This accommodates 450 nodes while remaining strictly contiguous and avoiding any overlap with the on-premises 10.20.2.0/24 network advertised over Cloud Interconnect or the cluster's Pod, Service, and control plane ranges.10.20.0.0/24 with 10.20.0.0/23 in the authorized networks list ensures both existing nodes and new nodes provisioned in 10.20.1.0/24 can reach the control plane endpoint.--master-authorized-networks replaces the entire existing allowlist rather than appending to it. Including both 192.168.10.0/24 and 192.168.20.0/24 alongside 10.20.0.0/23 ensures zero downtime for administrators while they transition, and removing 192.168.10.0/24 in a second update safely completes decommissioning.gcloud compute networks subnets expand-ip-range) happens live without draining nodes or recreating node pools./23 preserves clean routing across Cloud Interconnect without creating asymmetric or overlapping BGP prefixes.This approach is the only option that simultaneously provides enough non-overlapping IP addresses for 450 nodes, keeps newly scaled nodes authorized to communicate with the control plane, and performs a graceful cutover of administrative CIDR ranges without interrupting cluster operations.
In a Compute Engine VPC subnet, secondary IPv4 ranges (also known as alias IP ranges) are additional CIDR blocks attached to a subnet specifically so VPC-native GKE clusters can allocate IP addresses to Kubernetes Pods and Services separately from the host VM IP space.
10.20.4.0/23 as a secondary range only provides capacity for additional Pod CIDRs (via multi-Pod CIDR) or Services; the existing 10.20.0.0/24 primary range remains capped at 252 nodes.10.20.4.0/23 does not permit additional nodes to reach the control plane because nodes cannot be placed in that secondary range.ClusterIP Services.The cluster were running out of Pod IP addresses rather than node IP addresses—in which case adding a secondary Pod IPv4 range and registering it with --additional-pod-ipv4-ranges for a new node pool would be a valid scaling pattern.
This approach attempts to perform an in-place primary subnet expansion to a /22 prefix (10.20.0.0/22) and immediately switches the control plane authorized networks allowlist to the expanded node range and the new administrator CIDR (192.168.20.0/24).
10.20.0.0/22 CIDR block spans 10.20.0.0 through 10.20.3.255. This directly overlaps with the on-premises branch subnet 10.20.2.0/24 already reachable over Cloud Interconnect, introducing routing ambiguity and packet blackholing./22, the conflicting subnet cannot be reduced without tearing down the subnet and recreating the GKE cluster.192.168.10.0/24 before administrators have completed their migration to 192.168.20.0/24 immediately blocks existing administrators from accessing the Kubernetes API server.The 10.20.2.0/24 range were not used anywhere on-premises or in connected VPCs, and all administrators had already fully migrated off 192.168.10.0/24 prior to the authorized networks update.
This approach expands the primary node subnet to 10.20.0.0/23 and attempts to use the gcloud container clusters update command with --master-authorized-networks=192.168.20.0/24 under the assumption that GKE appends newly passed CIDR blocks to the existing allowlist.
--master-authorized-networks is declarative rather than additive. Passing only 192.168.20.0/24 replaces the entire allowlist, immediately revoking control plane access for both the current administrator range (192.168.10.0/24) and all cluster nodes (10.20.0.0/24).10.20.0.0/24 authorized after expanding the primary subnet to 10.20.0.0/23 leaves any new nodes assigned IPs in 10.20.1.0/24 unable to authenticate or register with the GKE control plane.192.168.20.0/24 before decommissioning 192.168.10.0/24.cidrBlocks.The GKE API supported an append-only flag for authorized networks and automatically synchronized primary subnet expansions into the control plane allowlist, neither of which occurs by default.
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