The most exciting feature that came out of the SQL Server 2012 release was AlwaysOn Availability Groups (AG). Even through AG has received a lot of publicity, there are many people who haven’t used it or may not even know what it means.
For the next couple of weeks, I am working on an AG project, so I thought it would be a good idea to blog about AG. Here are some facts:
- The AG feature is only available with Enterprise Edition of SQL Server 2012 and 2014.
- AG allows you to set up one or more replicas of a group of databases on other instances.
- SQL Server 2012 allows you to create up to four secondary replicas.
- SQL Server 2014 allows you to create up to eight secondary replicas.
- An AG group consists of one or more databases.
- You can configure secondary replicas to handle read-only workloads.
- You can offload full backups (copy-only) and transaction log backups onto a secondary.
- You can set up a secondary to be an automatic failover partner to accomplish high availability.
- AG does not require a SQL Server failover cluster, but it does require a Windows cluster.
- You can configure a name called a Listener to always point to the primary replica. Use this name in your connection strings.
- You can use the ApplicationIntent argument along with the Listener to redirect read-only workloads.
- Each replica must be configured either for synchronous or asynchronous commit mode. If configuring for synchronous commit mode, the transactions are hardened on the secondary before being committed on the primary.
- AG does not use shared storage. Each node will have a copy of the databases.
I hope this helps clarify the differences between failover clustering, database mirroring, and AG.