The official SQL Server AlwaysOn team blog.
Welcome to the official Microsoft SQL Server team blog for SQL Server AlwaysOn. As a team we are excited to share our knowledge with you and more importantly we are keen to get valuable feedback from you about the new high availability and disaster recovery solution for SQL Server, AlwaysOn.
Let us get started by answering the fundamental question, What is SQL Server AlwaysOn?
SQL Server AlwaysOn is the new high availability and disaster recovery solution for SQL Server. Using AlwaysOn businesses can achieve increased application availability for their mission critical applications and get higher returns on their high availability investments through better utlization of hardware resources. AlwaysOn also increases productivity and lowers TCO by greatly simplifying high availability deployment and management.
The new solution will be available in the next release of SQL Server, code named Denali. The initial feature set of AlwaysOn is now available in CTP1 that was just released during the SQL PASS 2010 conference. The CTP can be downloaded here.
Database Availability and Instance Availability using AlwaysOn
The key aspect of AlwaysOn is that it provides you with the flexibility to configure high availability and disaster recovery for your application databases and\or for the entire instance.
An important point to note is that AlwaysOn is built on the robust and reliable Windows Server Failover Cluster (WSFC) platform and is required for enabling Availability Group for database level availability and for creating a failover cluster instance for instance level availability. We will talk more about this dependency in the future blogs.
For getting started with AlwaysOn using CTP1, check the new AlwaysOn resource site. Note that you will see a lot of reference to HADR in the CTP1 download and documentation. This is our internal reference while the features are in development.
I'm wondering about the dependency on Windows Server Failover Cluster and the relation to Mirroring. It looks as though AlwaysOn Database Availability replaces SQL Mirroring (2008/2005) - does that mean that Windows Server Enterprise is now required for database failover? That would be a significant cost increase for many companies. Or is mirroring being left in, with DB-level HADR added for the higher end licenses?
If so, please don't forget about mirroring when working on the System Center modules for Denali - The lack of mirroring monitoring in SCOM is a complete nightmare right now!
Database Mirroring will be available in the next release as is. But it is quite possible that Availability Groups will replace database mirroring in subsequent releases. You are correct about the Windows Server enterprise SKU requirement. Availability Groups require the machines to be part of the same Windows Server Failover Cluster and the clustering feature is only supported in the enterprise edition.
With regard to the mom pack, thanks for the feedback. We will look into it.
We are investing in MOM pack for monitoring Availability Groups. So the experience will definitely be better in Denali for the new features.
recent Features is showing 2-node failover clustering available in SQL Server 2008 Standard Version. This seems to contradict previous statement. Can you clarify whether the standard edition will support a basic failover configuration or only the enterprise edition as stated above?
So you need to have a cluster to do this? You cannot do this from a Cluster to single server? Bah.. Total failure of a feature. One of the big failing point in a Cluster is Storage. So if your storage is down or working badly your cluster is down and so is your AlwaysOn feature.
And do not start to talk to me about doubling the Storage, it is done but the synchronization times in the case of a Controller failure are higher than one minute in a heavy IO environment and this will just bring your cluster on its knees, as well than your AlwaysOn feature.
Sorry but this is a so big disapointment..