HP ProLiant x64 Windows Unified Data Storage Server OS Upgrade T5322A Merkblatt

2

 

Introduction

 

Redundant Array of Independent Disks (RAID) is a widely-adopted control framework for aggregating 

the capacity and performance of hard disk drives, while also providing fault tolerance. When a hard 

disk failure occurs, data flow continues, uninterrupted, by deriving information from the remaining 

disks in the array. For environments that have direct-attached RAID, applications (clients) using the 
storage reside on the same computer; therefore, protecting only the disk devices is sufficient. 
In the consolidated storage model, the client that uses the storage no longer resides on the same 

computer, but rather on other computers on the network. The failure of the storage server in this 

environment has a much broader impact. To protect this storage resource, redundancy must now 
include other components of the system beyond the storage devices. Protecting the most vulnerable of 

these components (such as having multiple power supplies and NICs) is sufficient in most Small and 

Medium Business (SMB) environments; however, there are still some hardware components that 

cannot be made redundant. 
To minimize the disruption window further, users can use software clustering. Software clustering is 

the ability to configure multiple servers to work as a single unit. Clustering protects the storage system 

from the unlikely event of failures caused by unprotected hardware. In this configuration, the failure of 

one server does not lead to total system outage. Instead, when one server fails, the workload shifts to 
the second server in the cluster, and work proceeds as normal. 
Cluster environments that use external SCSI or Fibre Channel (FC) storage arrays that have built-in 

redundancy (multiple disk controllers, power supply, and so on) are expensive. Due to its high cost, 

this storage is reserved for the small number of highly critical clients. Administrators can maximize the 

usage of this redundant storage resource at a much lower per-client cost and achieve virtually 
continuous uptime by leveraging the functionalities of Microsoft® Cluster Service (MSCS) and 

Microsoft iSCSI Software Target. This document details the steps to configure Microsoft iSCSI 

Software Target for high availability using MSCS. The current release of Microsoft iSCSI Software 

Target only supports two-node clusters in an active/passive configuration.  

Figure 1. Mcrosoft iSCSI Software Target and MSCS provide redundancy for iSCSI disks.