Examkin

Soru

Zorluk: OrtaResilience, High Availability, and Redundancy

A cloud security architect is evaluating a high-availability strategy for a critical online payment application. The application requires near-zero Recovery Time Objective (RTO) and real-time transaction consistency across two distinct cloud regions. A system administrator proposes using asynchronous database replication paired with automated failover via DNS routing, but without implementing a third-site witness or quorum node. Which of the following risks is MOST likely to occur if a network partition isolates the primary region while servers in both regions remain fully operational?

  1. A split-brain scenario where database nodes in both regions assume the primary role and accept writes concurrently, causing severe data inconsistency.Cevap
  2. B
    Complete loss of hardware fault tolerance due to local disk RAID array corruption within the primary data center.
  3. C
    Misclassification of DNS failover routing as a detective security control rather than an active-passive availability control.
  4. D
    Inability of the secondary database instance to preserve non-repudiation logs due to missing digital signatures across redundant nodes.

Cevap

A split-brain scenario where database nodes in both regions assume the primary role and accept writes concurrently, causing severe data inconsistency.
In high-availability multi-region cluster design, a split-brain condition occurs when a network partition breaks communication between nodes while both remain powered on. Without a third-site witness or odd-numbered quorum cluster to establish majority consensus, both nodes believe the other has failed and simultaneously assume the active role, accepting conflicting data writes.

Adım Adım Çözüm

1
Analyze the architecture components and failure condition.
Two cloud regions are connected via asynchronous replication without a quorum/witness server, and a network partition breaks inter-region communication while both sites stay online.
When the heartbeat/replication link between two isolated sites drops, each site must determine whether the other site is down or simply unreachable.
2
Evaluate how active-passive high availability systems behave without a quorum mechanism.
Because neither site can communicate with the other and there is no third-party witness to break the tie, both nodes assume the peer is dead.
Without quorum or majority voting, both nodes promote themselves to active status.
3
Determine the impact of dual active nodes accepting concurrent writes.
Both instances accept user transactions independently, creating divergent database states that lead to data corruption when connectivity is restored (split-brain condition).
Split-brain is the primary high-availability failure mode when heartbeats fail across isolated active-passive nodes without quorum.

Anahtar Kavram

High Availability Quorum and Split-Brain Conditions
Tahmini Süre:1m 30s
Bu soruyu puanla

Konular

Resilience, High Availability, and Redundancy

Bu konudan daha fazla soru