2+1 Scenario Witness Deployment Guide
Read the full document on Sangfor Cloud Platform (SCP)
1. Introduction & The Core Challenge
Comprehensive Guide to Sangfor HCI 2+1 Witness Server Deployment: Preventing Data inconsistency and Ensuring Data Integrity In the realm of Hyper-Converged Infrastructure (HCI), maintaining data consistency and preventing catastrophic scenarios like "Data inconsistency" are critical priorities. For organizations deploying a two-node HCI cluster, the risk of data inconsistency during a storage network failure is a significant concern.
2. Guide Overview & Target Audience
The Sangfor HCI 2+1 Scenario Witness Deployment Guide provides a vital solution to this challenge, detailing the deployment and configuration of a witness server to ensure robust high availability and data protection. This comprehensive summary explores the essential concepts, deployment requirements, and implementation processes outlined in the guide, serving as a valuable resource for system administrators and O&M personnel.
3. Mechanism & Mitigation Strategy
The Threat of Data inconsistency in Two-Node Clusters The primary motivation for deploying a 2+1 witness architecture stems from the inherent vulnerabilities of a standard two-node HCI cluster. In such a setup, if the storage network experiences an abnormality or failure, the two nodes may lose communication with each other while still remaining active.. The result is severe data inconsistency, with multiple replicas diverging and no clear "source of truth." To combat this, Sangfor introduced the witness mechanism, a critical feature available from HCI version 5.2 onwards. The witness mechanism acts as an impartial third party. When network interruptions threaten data consistency, the witness helps the HCI cluster determine the correct data replica, effectively preventing Data inconsistency scenarios and ensuring that virtual machines operate only on nodes with valid, synchronized data.