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 runs a private GKE cluster in us-central1 inside vpc-gke with a private control plane endpoint (172.16.0.16/28) and control plane global access enabled. Private kubectl access is required from two networks already listed in Control Plane Authorized Networks:
10.200.0.0/16): Connected to vpc-gke via Dedicated Interconnect and a Cloud Router in us-east4 (vpc-gke uses regional dynamic routing and default BGP advertisements).10.50.0.0/24): Hosted in vpc-mgmt, which is connected to vpc-gke via VPC Network Peering.Neither environment can currently reach 172.16.0.16/28.
Which configuration changes should you implement?
vpc-mgmt and vpc-gke, enable Export custom routes and Import custom routes on both the vpc-mgmt peering and the GKE-managed peering, and configure custom BGP route advertisements for 172.16.0.16/28 on the Interconnect Cloud Router while retaining regional dynamic routing.This configuration attempts to use VPC Network Peering custom route exchange to propagate routes between vpc-mgmt, vpc-gke, and the Google-managed GKE control plane VPC, while adding a custom Border Gateway Protocol (BGP) advertisement for 172.16.0.16/28 on the us-east4 Interconnect Cloud Router and keeping regional dynamic routing mode.
vpc-gke. Keeping VPC Network Peering between vpc-mgmt and vpc-gke creates a two-hop peering chain (vpc-mgmt <-> vpc-gke <-> GKE control plane VPC). Google Cloud strictly prohibits transitive routing across multiple peering connections, so bastion hosts in vpc-mgmt still cannot reach 172.16.0.16/28.vpc-gke in regional dynamic routing mode, routes learned by the Dedicated Interconnect Cloud Router in us-east4 (10.200.0.0/16) are only programmed in us-east4. The GKE control plane peering in us-central1 cannot export 10.200.0.0/16 to the control plane VPC, causing return traffic from 172.16.0.16/28 to drop.all_subnets advertisement.vpc-gke.This approach would only work partially if the bastion hosts were moved directly into vpc-gke (eliminating the second peering hop) and vpc-gke were switched to global dynamic routing mode.
vpc-gke to global dynamic routing mode, enable Export custom routes on the GKE-managed VPC peering, and replace the vpc-mgmt peering with HA VPN while leaving both the Interconnect and HA VPN Cloud Routers in default route advertisement mode.This option correctly establishes global dynamic routing mode on vpc-gke, enables Export custom routes on the GKE-managed peering, and replaces the vpc-mgmt VPC peering with HA VPN. However, it leaves both the Dedicated Interconnect Cloud Router and the HA VPN Cloud Router in default route advertisement mode.
172.16.0.16/28 on-premises and in vpc-mgmt: In default advertisement mode, Cloud Router only advertises local VPC subnet routes (all_subnets when global dynamic routing is enabled). Because 172.16.0.16/28 is a subnet inside a Google-managed peered VPC rather than a native subnet of vpc-gke, neither Cloud Router advertises it over BGP.vpc-gke learns 10.200.0.0/16 and 10.50.0.0/24 and exports them to the control plane VPC, clients on-premises and on the bastion hosts have no BGP route to 172.16.0.16/28, so kubectl packets never leave their source networks toward vpc-gke.all_subnets alongside the custom range (172.16.0.16/28) so existing VPC subnets continue to be advertised.This configuration would work if both Cloud Routers in vpc-gke were updated to custom route advertisement mode configured to announce both all_subnets and 172.16.0.16/28.
vpc-gke to global dynamic routing mode, connect vpc-mgmt and vpc-gke as VPC spokes in a Network Connectivity Center (NCC) hub, and configure custom BGP route advertisements for 172.16.0.16/28 on the Interconnect Cloud Router without modifying the GKE-managed VPC peering.This option enables global dynamic routing on vpc-gke, advertises 172.16.0.16/28 from the Interconnect Cloud Router, uses Network Connectivity Center (NCC) VPC spokes to link vpc-mgmt and vpc-gke, and leaves the GKE-managed VPC peering connection unmodified.
vpc-gke, the Google-managed control plane VPC never imports 10.200.0.0/16 learned over Dedicated Interconnect. Inbound kubectl requests from on-premises reach 172.16.0.16/28, but the control plane drops all return packets.vpc-mgmt still lack reachability to 172.16.0.16/28.vpc-mgmt and vpc-gke is specifically required when vpc-mgmt must reach endpoints hosted behind a VPC-peered network in vpc-gke.This approach would be valid for reaching worker node pods or Private Service Connect endpoints inside vpc-gke if vpc-mgmt were not targeting a VPC-peered control plane, and if Export custom routes were enabled on the GKE peering for the Interconnect path.
vpc-gke to global dynamic routing mode, enable Export custom routes on the GKE-managed VPC peering, replace the vpc-mgmt peering with HA VPN using Cloud Routers, and configure custom BGP route advertisements for 172.16.0.16/28 alongside all_subnets on both the Interconnect and HA VPN Cloud Routers in vpc-gke.This architecture combines global dynamic routing, custom Cloud Router BGP advertisements, VPC Network Peering custom route export, and HA VPN inter-VPC connectivity to establish end-to-end private reachability to the GKE control plane from both on-premises and vpc-mgmt.
vpc-gke to global dynamic routing mode allows the us-east4 Interconnect Cloud Router to advertise subnets across all regions (all_subnets) and makes learned on-premises routes (10.200.0.0/16) available in us-central1.172.16.0.16/28 plus all_subnets) on the Interconnect Cloud Router propagates the GKE control plane /28 CIDR to on-premises routers, because Cloud Router does not automatically advertise VPC-peered ranges.vpc-gke lets the Google-managed control plane VPC import 10.200.0.0/16 for return traffic.vpc-mgmt:
vpc-mgmt and vpc-gke eliminates the two-hop peering chain. Packets from vpc-mgmt arrive over the HA VPN tunnel directly into vpc-gke and cross only one peering hop to the GKE control plane.172.16.0.16/28 on the vpc-gke HA VPN Cloud Router teaches vpc-mgmt the control plane route, while vpc-mgmt advertises 10.50.0.0/24 back to vpc-gke so the GKE peering exports it to the control plane.vpc-mgmt and vpc-gke with BGP-routed HA VPN tunnels.Because GKE private control planes sit behind a Google-managed VPC Network Peering boundary, no multi-hop peering topology can provide direct Layer 3 reachability. Terminating both external (Interconnect) and inter-VPC (HA VPN) BGP sessions directly in vpc-gke with custom route advertisements and peering custom route export is the standard, fully managed Google Cloud pattern.
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