Examkin

Question

Difficulty: HardDeploying and Managing Google Kubernetes Engine (GKE) Clusters and Workloads

A data engineering team is preparing to deploy an internal stateful batch processing application to Google Kubernetes Engine (GKE). The application requires specific Linux kernel parameter modifications via sysctl settings on the underlying node OS to achieve optimal network socket throughput. Furthermore, the batch processing job cannot tolerate sudden node terminations or interruptions during its multi-hour execution window. Which deployment configuration should the cloud engineer implement to satisfy these technical requirements?

  1. Provision a GKE Standard cluster with a node pool composed of standard (on-demand) Compute Engine virtual machines, and configure the required sysctl kernel parameters on the node pool.Answer
  2. B
    Provision a GKE Autopilot cluster and specify the custom sysctl kernel parameters directly inside the Pod deployment specification under securityContext.
  3. C
    Provision a GKE Standard cluster with a Spot VM node pool to minimize compute costs, relying on Kubernetes StatefulSet persistent volumes to recover from node preemptions.
  4. D
    Provision a GKE Autopilot cluster and execute `kubectl config set-cluster` to elevate cluster administrator privileges for modifying worker node OS settings.

Answer

Provision a GKE Standard cluster using a standard on-demand node pool where sysctl kernel tuneables can be configured.
GKE Standard provides full operational control over node pools, enabling custom Linux sysctl kernel tuning needed for specific network performance profile requirements. Additionally, standard on-demand virtual machines ensure that long-running stateful batch workloads run without unexpected termination.

Step-by-Step Solution

1
Evaluate cluster mode operational boundaries
GKE Autopilot locks down underlying node OS configurations and restricts custom sysctl kernel parameters, whereas GKE Standard supports custom node configuration.
The workload specifically requires custom sysctl kernel modifications on the host node OS.
2
Evaluate node provisioning type for workload fault tolerance
Spot VMs are subject to sudden preemption, which breaks uninterrupted multi-hour execution required for this stateful workload.
Standard on-demand VMs guarantee continuous execution without preemption risk.
3
Select optimal combination
GKE Standard with standard on-demand Compute Engine node pool meets both the sysctl customization and zero-preemption requirements.
This configuration satisfies all technical constraints while keeping infrastructure reliable.

Key Concept

GKE Standard vs Autopilot cluster capabilities and node pool VM instance suitability
Estimated Time:2m 0s
Rate this question

Topics

Deploying and Managing Google Kubernetes Engine (GKE) Clusters and Workloads