Forum Discussion
General Use File server on Failover Cluster deployed in Azure on Windows server 2016
I am trying to get a clear answer on Failover clustering in WIndows Server 2016. In Windows 2012 setting up a Scale Out File Server share was to be used for SQL or Hyper-V. Using the General Use File server was for file shares. I would like to know if this guidance is still true for WIndows 2016? ON this blog https://docs.microsoft.com/en-ca/windows-server/remote/remote-desktop-services/rds-storage-spaces-direct-deployment Microsoft is recommending a SOFS for user profile disks. Doesn't that go agains the guidance Microsoft recommends?
In the end I want to build a HA file share to store citrix user profiles in Azure. I am not sure what to use to accomplish it in Azure, a SOFS or a gerneral use file share?
It is still true.
SOFS in S2D (Storage Spaces Direct) or in a disaggregated setting (2+ nodes + JBOD(s)) is meant for storaging VHDX files that are used in a Hyper-V compute cluster.
A simple cluster setup that can host a Highly Available file server VM would be an asymmetric cluster:
- (2) Nodes with (2) SAS HBAs
- (1) JBOD or more if needed
The above setup would host both clustered Storage Spaces and Hyper-V. An example on our company blog.
It is still true.
SOFS in S2D (Storage Spaces Direct) or in a disaggregated setting (2+ nodes + JBOD(s)) is meant for storaging VHDX files that are used in a Hyper-V compute cluster.
A simple cluster setup that can host a Highly Available file server VM would be an asymmetric cluster:
- (2) Nodes with (2) SAS HBAs
- (1) JBOD or more if needed
The above setup would host both clustered Storage Spaces and Hyper-V. An example on our company blog.
- Dennis PeabodyCopper Contributor
Hey Bradley,
That is exactly what I am trying to do. The SOFS was easy to setup, but not ideal for Citrix Profiles as it stores them in indivdual files (unlike RDS, which is VHD and would work).
I tried setting up a general file server in the failover cluster, but I can only get it to reply when it is on one of the two nodes. It must have something to do with the load balancer, but I have not worked it out yet.
Did you come up with a solution for this?