SQL Server Clustering technology

iSCSI SAN Storage Software / Application Solutions / Microsoft SQL Server / SQL Server Clustering technology

SQL Server Clustering technology

SQL Server clustering is a high-availability technology that helps companies meet their availability and uptime goals. This technology ensures that mission-critical services, applications, and data will be constantly available in case of any kind of failure or disaster. Being an automatic technology, it allows one physical server to take control over tasks and responsibilities of another one that failed. SQL Server cluster shares tasks and services between cluster nodes, each of which can contain one or more shared disks grouped into logical units called resource groups.

Basics of Clustering

Every virtual server is shown on a network as a system. If a virtual server contains SQL Server resources, clients connected to it may access resources on its current host node. If the current host fails, the resource group will be transferred to another node within the cluster. It is critical to set failover order for each instance in clusters having more than two nodes or instances.

The aim of clustering is to provide availability to clients by supplying a system with an automatic failover mechanism. During the failure there will be a short database server interruption on all clusters. Clustering is a part of a big strategy to help reduce application downtime.

Clustering Topologies

There are four types of cluster topologies:

• Single instance (active-passive)
• Multiple instance (active-active)
• N+1
• N+M

Single instance: In such kind of clusters one node owns all resource groups, while the other nodes are held standby. Should the primary node fail, the resource groups would be transferred to the standby node. When the standby node comes online, it will typically consist of disks containing the database files and transaction logs. When the secondary node becomes active, SQL Server starts up on the virtual server and recovers the database. This topology is called active-passive.

Multiple instances: In such kind of a cluster (active-active) resource groups are shared between virtual servers. In multiple-instance cluster each node can host two or more virtual servers at the same time. Clusters of such topology are more cost effective than single-instance clusters because there are no nodes on hold or standby. The multiple instance clusters are used when high performance and load balancing is required.

Multiple instances: In such kind of a cluster (active-active) resource groups are shared between virtual servers. In multiple-instance cluster each node can host two or more virtual servers at the same time. Clusters of such topology are more cost effective than single-instance clusters because there are no nodes on hold or standby. The multiple instance clusters are used when high performance and load balancing is required.

N+1: This topology is modification of multiple-instance clustering where two or more nodes share the same failover node. The standby node requires significant hardware capabilities to support all N servers for the situations when they all fail simultaneously. N+1 uses resources effectively because there is only one standby node offline. However, depending on the hardware it does not offer the performance of multiple-instance clustering.

N+M: In such kind of clusters there are two or more working nodes and two or more standby nodes. Such configuration offers costs-effectiveness of N+1 having at the same time lower chance of performance degradation during a multiple failure as the failover node or nodes do not have to support all failed nodes.

Clustering Requirements

Network requirements: Clustering requires a private network that will be used by all nodes in a cluster. Moreover, clustering services use private communication channel to keep in touch with each other. As a matter of fact, it allows the cluster to act properly even if the public network is offline.

Hardware requirements: For example, cluster that includes two SQL Servers will need the following hardware:

• Two RAID arrays that support RAID 5 or RAID 10
• Two servers of a minimum 1GB RAM and a single Intel Xeon CPU
• Each server must have at least one local SATA or SAS hard disk
• Each server must have a iSCSI or Fiber Channel adapter in order to communicate with the shared disk array
• Each server must at least have two networks cards.

Software requirements:

• Microsoft Windows 2008R2 Datacenter Licenses
• One SQL 2008 R2 Standard
• The latest Service Packs.

Conclusion

Nowadays clustering on SQL Server is a reliable technology, but it still has its pros and cons. In order to avoid unnecessary failures, one should organize operating processes and procedures to work specifically with SQL Server cluster. Make sure that you address all of the clustering dependencies to ensure high availability with SQL Server clustering.

Download SQL Server Clustering Guide:

---

Live demo		Free 30-day trial		How to buy
Request now		Download free trial		Request a quote