The AlwaysOn High Availability Groups technology is intended to replace database mirror- ing. It has all the benefits of database mirroring plus many more. Because many of the issues with hardware limitations have been addressed, it’s a far more robust product than what was available in SQL Server 2008. Database mirroring is still available in SQL Server 2012, but it’s being phased out over the next few releases. Some new features of AlwaysOn High Availabil- ity Groups are as follows:
■
■ It can fail over a whole group of databases instead of just one at a time. ■
■ It uses a virtual name to provide faster failover. ■
■ Multiple secondary servers are supported, up to four. ■
■ Asynchronous active log synchronization is available for solutions that require high
latency.
■
■ Built-in encryption and compression provide more secure and faster data transfer. ■
■ Automatic page repair is available for corrupted pages. ■
■ It provides a flexible failover policy with more control. ■
■ It uses two synchronous secondary servers.
As you can see, quite a few features have been added to database mirroring. From this version forward, it’s the recommended solution for high availability.
EXAM TIP
A database can use AlwaysOn high Availability Groups or database mirroring, but not both. Using AlwaysOn High Availability Groups is a much more involved solution to implement and requires considerable configuration both on SQL Server as well as the Windows Server that it’s installed on. You need to know the prerequisites involved when deciding whether AlwaysOn should be part of the SharePoint farm’s high-availability solution. The prerequisites for configuring AlwaysOn High Availability Groups are as follows:
■
■ A server with AlwaysOn installed can’t be a domain controller. ■
■ It can’t be installed on a server with WOW64 (Windows 32-bit on a 64-bit server). ■
■ Each server in the group has to be a node in a WSFC cluster. ■
■ The WSFC cluster must contain sufficient nodes to support the availability group con-
figurations.
■
■ All applicable Windows hotfixes have been installed on every node in the WSFC cluster.
MORE INFO ALWAYSON HIGH AVAILABILITY GROUPS
See http://msdn.microsoft.com/library/ff878487(v=sql.110).aspx for more information on prerequisites, restrictions, and recommendations for AlwaysOn High Availability Groups.
WSFC is an advanced topic, but a general understanding of it is important to grasp the concepts of AlwaysOn High Availability Groups as well as AlwaysOn Failover Cluster Instances. The term clustering describes several different technologies used in planning and configuring SharePoint farms. Clearly stating which type of clustering is being used helps avoid confusion during the planning process.
WSFC is a Windows Server technology that supports high availability and disaster recov- ery for applications that sit on top of it, such as SQL Server and Exchange Server. Whenever a node in a cluster fails or is turned off for maintenance, it can automatically (or manually) transfer those services to another node in the cluster. In SQL Server, an AlwaysOn high- availability group becomes a WSFC cluster resource. WSFC then monitors the health of each node and initiates the failover processes. Although knowing every detail of how to configure a WSFC instance isn’t necessary, understanding the terms helps with an overall comprehen- sion of high-availability options. Some terms used with WSFC are as follows:
■
■ WSFC cluster A group of independent servers that work together to provide high
availability of services and applications such as SQL Server, Exchange, and/or Share- Point.
■
■ Failover cluster instance A Windows Server service that manages an IP resource,
a network name resource, and the application and/or services to be managed by the cluster. Rather than connect to an IP or computer name, an application connects to the IP resource or network name resource that crosses over all servers in the cluster.
■
■ Node A server that’s part of the cluster. ■
■ IP resource An IP address (or IPv6 address) that refers to the cluster as a single
entity.
■
■ Network name resource Used in place of a server name or network name so that
applications and servers can refer to the cluster as a whole.
■
■ Resource group The set of resources typically required to manage an application or
service as a single entity. Failover and fallback are done as a resource group.
■
■ Cluster resource An object that can be “owned” by a node. It can be made online,
offline, or transferred by a node but not shared by nodes.
■
■ Preferred owner The node in a cluster that a resource group prefers to run on. ■
■ Possible owner A node that a resource group can run on if the preferred owner is
unavailable.
You can see, in these terms, many similarities between “typical” SQL Server clustering and WSFC. AlwaysOn High Availability Groups might seem to have more in common with tradi- tional clustering than with database mirroring. In reality, the technology has some aspects of both, plus additional enhancements. One way to think about AlwaysOn high availability is that it works with WSFC to tie several resources together—database name, IP resource, net- work name resource, and SQL Server instance—so that they can act as a single entity regard- less of what node is actively providing them.
MORE INFO WINDOWS SERVER FAILOVER CLUSTERING WITH SQL SERVER
See http://msdn.microsoft.com/en-US/library/hh270278.aspx for more information on us- ing WSFC with SQL Server.