SQL Cluster without a shared disk. Possible?

WePython 1 Months+

Is it possible to create a two node cluster using two basic network servers that only have local disks?

One will act as an idle standby server. Without knowing too much about clustering I'm not sure if a separate shared disk array is required.

If so, is it possible to use a DAS setup, or does it have to be a SAN that supports clustering?


I am afraid that this is not possible. Clustering works on the basis of resource ownership and acquisition; so anything that is not accessible to all nodes in the cluster cannot be used.

If shared storage is not an option for you then have a look at using mirroring which does not require storage to be shared across multiple servers. You could even use a witness server to automatically fail over should the primary become unreachable for any reason:





I hope this helps you.


This is sort of possible using the SQL 2012 AlwaysOn feature, which allows you to create a cluster between two nodes that do not share storage. The setup uses database mirroring behind the scenes to replicate the data to the remote nodes storage but I don't think automatic failover is possible, would have to be kicked off by a user or process.


No, it is not possible to create a SQL cluster as you have described it.


Not in 2008, but in SQL 2012 you don't need to have shared storage and it WILL perform automatic failover if configured, as well as seperating read and write load, offsite backups and automatic corruption detection and repair.

You only need to file structure (disk letters and folder locations to be identical on the servers) if your going to consider this... check out these links...


then use brent's great checklist (this also contains further links to important pre reqs that you MUST perform and test before installing AlwaysON :


the only showstopper may be, this is only available in enterprise edition...

Previous : Single or multiple databases?
Next : Default Constraints with Merge Replication