Examkin

Soru

Zorluk: OrtaPlanning Google Kubernetes Engine (GKE) Cluster Architectures

An enterprise financial organization is architecting a Google Kubernetes Engine (GKE) cluster to support a mission-critical transaction processing system. The architecture must ensure high availability for both the cluster control plane and worker nodes during a single availability zone outage, while avoiding control plane downtime during master upgrades. Additionally, the workload cannot tolerate unexpected node preemptions. Which cluster configuration should the cloud engineer choose to satisfy these constraints?

  1. Provision a Regional GKE cluster distributed across multiple availability zones using standard persistent node pools.Cevap
  2. B
    Provision a Zonal GKE cluster with multi-zone node pool locations and automated node upgrades.
  3. C
    Provision a Zonal GKE Autopilot cluster configured to automatically replicate pods across multiple GCP regions.
  4. D
    Provision a Regional GKE cluster using Spot VM node pools for all transaction processing application nodes.

Cevap

Provision a Regional GKE cluster distributed across multiple availability zones using standard persistent node pools.
Regional GKE clusters replicate the control plane across three availability zones within the chosen region and spread node pools across those zones. This guarantees high availability during zone outages and allows master upgrades without control plane downtime. Using standard (non-Spot) nodes ensures worker instances are not unexpectedly preempted.

Adım Adım Çözüm

1
Evaluate control plane availability requirements
Zonal GKE clusters deploy a single master node in one zone, whereas Regional clusters deploy three master instances across three zones.
High availability for control plane upgrades and zone outages requires a Regional cluster architecture.
2
Evaluate workload fault tolerance against VM preemption
Spot VMs can be reclaimed by Google Compute Engine at any time with a 30-second notice.
Mission-critical transaction processing cannot tolerate unexpected node preemption, ruling out Spot VM node pools.
3
Select the GKE cluster topology that fulfills all requirements
A Regional GKE cluster with standard (non-preemptible) worker nodes distributed across multiple zones fulfills both high availability and persistent node requirements.
This configuration balances control plane resilience and worker node stability.

Anahtar Kavram

GKE Regional vs. Zonal Cluster Architecture and Spot VM Suitability
Bu soruyu puanla

Konular

Planning Google Kubernetes Engine (GKE) Cluster Architectures

Bu konudan daha fazla soru