Examkin

Soru

Zorluk: ZorResilience, High Availability, and Redundancy

A security engineer is configuring an automated failover workflow for an active-passive high-availability database cluster to prevent split-brain conditions and ensure data integrity during an ungraceful primary node failure. Arrange the operational steps in the correct chronological order from initial failure detection to full service restoration on the standby node.

  1. 1Detect primary node heartbeat loss and confirm the failure threshold timeout.
  2. 2Execute a STONITH (Shoot The Other Node In The Head) fencing command against the failed primary node.
  3. 3Verify cluster quorum consensus and formally promote the standby node to the active role.
  4. 4Promote the underlying storage volume to read-write mode and replay uncommitted transaction logs.
  5. 5Rebind the cluster Virtual IP (VIP) address to the secondary node's network interface controller.

Cevap

The correct chronological failover sequence is: 1) Detect primary node heartbeat loss and confirm the failure threshold timeout, 2) Execute a STONITH fencing command against the failed primary node, 3) Verify cluster quorum consensus and formally promote the standby node to the active role, 4) Promote the underlying storage volume to read-write mode and replay uncommitted transaction logs, and 5) Rebind the cluster Virtual IP (VIP) address to the secondary node's network interface controller.
In high-availability failover architectures, maintaining data integrity during ungraceful failures requires a strict order of operations: first, health monitors confirm heartbeat loss beyond the tolerance threshold; second, node fencing (STONITH) forcefully isolates the unresponsive primary to prevent split-brain scenario; third, surviving nodes verify quorum and promote the standby server; fourth, storage volumes are transitioned to read-write state and logs are replayed for crash recovery; fifth, the Virtual IP (VIP) is reassigned to the new primary so application traffic resumes transparently.

Adım Adım Çözüm

1
Identify initial failure condition.
Heartbeat failure is confirmed after the configured timer expires.
Prevents premature failover actions caused by transient network spikes.
2
Enforce cluster node fencing (STONITH).
The failed node is physically powered off or isolated via PDU/IPMI controls.
Fencing eliminates the possibility of split-brain conditions where two nodes attempt to access and corrupt shared storage concurrently.
3
Validate cluster quorum and promote node status.
The standby node transitions from secondary to primary status.
Ensures that failover decisions are authorized by a surviving cluster majority.
4
Mount and validate persistent storage state.
Storage volume becomes writable and database state is consistent.
Database operations must achieve crash consistency prior to receiving incoming application queries.
5
Migrate client network traffic.
Gratuitous ARP broadcasts or API calls shift the Virtual IP (VIP) to the new active node.
Reroutes application connections to the newly promoted active server with minimal client re-configuration.

Anahtar Kavram

High Availability Cluster Automated Failover and Fencing (STONITH) Sequence
Tahmini Süre:2m 30s
Bu soruyu puanla

Konular

Resilience, High Availability, and Redundancy

Bu konudan daha fazla soru