SQL Server 2012 Availability Enhancements Part I

Availability is one of the areas where Microsoft have felt a bit behind the competition. In this release, huge amounts of resources have been spent to catch up.

Although i’m about sum it up in 3 short paragraphs, don’t let that fool you. These are major revisions to the product and are of the highest importance. If you ever needed one of those, you’ll know it.

“AlwaysOn” is the name Microsoft have chosen for the new feature set of High Availability and Disaster Recovery (HADR). Although originally called HADRON, after the first few spelling mistakes where some of the letters got mixed up (you find it out…), the name was changed…

AlwaysOn Failover Cluster Instances. 

Failover clusters have been available for many years, but in this release 2 major, long waited enhancements have been added.

First, is the option to configure nodes on multiple subnets. These subents can be in the same physical location or spread across multiple dispersed geographical locations (Geo-Cluster, AKA Stretch Cluster). Of course, since you cannot expect a shared storage to support all nodes, this works using replication of data. Having multiple copies of the data also serves as a DR solution and not just HA. For more info, see  BOL topic “SQL Server Multi-Subnet Clustering”

Another major enhancement to the clustering abilities of SQL Server is the “Flexible Failover Policy“. If you’ve ever suffered the results a failover due to ‘minor’ issues (for example, network congestion) you’ll appreciate this one. With this new ability, YOU will be able to control how the cluster responds to failure of individual resources, in what order should the resources be taken offline, time out settings and much more.

In Part II, I will discuss AlwaysOn Availability Groups.

*NOTE: The features described here for SQL Server 2012 are available in CTP3 (Community Technology Preview) version and may or may not make it to the RTM (Release To Manufacture) version. It has happened before…