Specifications
Failover Clustering with Windows Server 2008 R2 CHAPTER 4 65
including Internet Information Services (IIS), Cluster Server, SQL Server 7.0 Enterprise Edition,
Microsoft Distributed Transaction Coordinator (MSDTC) 2.0, and sometimes the Windows NT
4.0 Option Pack. Moreover, the hardware support, driver support, and documentation were
not as forthcoming as they are today. Many IT organizations came to believe that failover
clustering was a difcult technology to install and maintain. That has all changed, thanks to the
efforts of the SQL Server and Failover Clustering product groups at Microsoft. Today, forming a
cluster with SQL Server 2008 R2 on Windows Server 2008 R2 is very easy. In addition, the two
technologies combined provide maximum availability compared to previous versions, espe-
cially for database administrators who want to virtualize their SQL Server workloads.
Now that you know some of the history behind failover clustering, it's time to take a
closer look into what failover clustering is all about and what it means for organizations and
database administrators. A SQL Server failover cluster is built on the foundation of a Windows
failover cluster, while providing high availability and protecting the whole instance of SQL
Server in the event of a server failure. Failover clustering allows organizations to meet their
high availability uptime requirements through redundancy in their SQL Server infrastructure
by eliminating single points of failure for the clustered application. The server that is used to
form a cluster can be either physical or virtual. The next section introduces the different types
of failover clusters that can be achieved with these two products (SQL Server 2008 R2 and
Windows Server 2008 R2), which work very well with one another.
Traditional Failover Clustering
The traditional SQL Server failover cluster has been around for years. With a traditional
failover cluster, there are two or more nodes (servers) connected to shared storage. A quorum
is formed between all nodes in the failover cluster, and this quorum determines the health
and number of failures the failover cluster can sustain. Communication between cluster nodes
is required for cluster operations and is achieved by using two or more independent networks
that connect the nodes of a cluster to avoid a single point of failure. SQL Server 2008 R2 is
installed on all nodes within a failover cluster. If a node in the cluster fails, the SQL Server
instance automatically fails over to a surviving node within the failover cluster. Note that the
failover is seamless from an end-user or application perspective. Like its predecessor, SQL
Server 2008 R2 delivers single-instance and multiple-instance failover cluster congurations.
In addition, SQL Server 2008 R2 on Windows Server 2008 R2 supports up to 16 nodes and a
maximum of 23 instances within a failover cluster due to the drive letter limitation.
IMPORTANT When you are conguring a cluster, make sure to connect the nodes by
more than one network; otherwise Microsoft Product Support Services does not support
the implementation. In addition, it is a best practice to always use more than one network.