Shared External SCSI Storage Cluster Configuration

SCSI Storage Clustering Configuration Consulting Services

In a cluster, shared disk storage is used to hold application data and two quorum partitions. Because this storage must be available to both cluster systems, it cannot be located on disks that depend on the availability of any one system. There are two methods you can use to implement shared access to disk storage. A single-initiator configuration enables you to use a private, single-initiator SCSI bus or FC interconnect to connect a cluster system to a storage enclosure. This configuration requires a storage enclosure with multiple, independent host channels. In addition, the enclosure must provide simultaneous access to the storage units from each host channel.

For the latest information about Shared External SCSI Storage Cluster Configuration

Network Load Balancing with External SCSI Storage

To provide redundancy, high availability, and scalability to the front-end Web services, we use the Network Load Balancing feature of Windows 2000 to build a cluster of four Web servers. This clustering feature allows multiple servers to work as a single system, thereby allowing failover and load balancing among the servers. In the back-end network, the two identical database servers are connected to a shared external SCSI storage device. We will install and configure the Cluster service of Windows 2000 and the failover clustering of SQL Server 2000 on these servers to form a database server cluster.

For the latest information about Network Load Balancing with External SCSI Storage

RAID SCSI Clustering

Each database server is installed with a dedicated redundant array of independent disks (RAID) SCSI controller card that in turn is connected to the shared external SCSI storage device. For more information about RAID technology, refer to the reference section at the end of this paper. Each of the servers has three network interface cards (NICs). One of the NICs is used for private cluster communication between the two servers. This NIC provides the cluster "heartbeat mechanism" and is connected to the other cluster server through an Ethernet crossover cable. One of the other NICs is connected to the back-end network, providing access to the front-end Web servers through a 100 megabits per second (Mbps) network switch. The remaining NIC is connected to a back-end management network where server administration is performed. This NIC also acts as a redundant backup connection for private cluster communication.

For the latest information about RAID SCSI Clustering