This may be a dumb question but I am looking to determine whether or not windows server 2012R2 naitively supports Mirroring of CSV's. We have a client using Datacore who has two P2000 SAN's. The Datacore boxes are set up in a cluster like configuration
and handle the storage layer all the virtual machine live on. Datacore is configured to mirror one SAN to the other and should allow for resources to function normally in the event one SAN dies. They would like to move away from Datacore and use native
windows server capabilities. The ideal scenario I am trying to achieve is converting the two Datacore boxes into scale out file servers, connecting each SOFS to each SAN and creating two CSV's (One for Each SAN), lastly serving the storage to the hyper-v servers
and VMM through a continuously available SMB3 share. I have done this in a lab network but i can't seem to figure out how to, or if its even possible to, provide the client with the same level of redundancy they are achieving now. In order to do
this I would need to then mirror one CSV (Or SAN) to the other CSV to protect again enclosure failure (Keep in mind i cannot use storage spaces and would need to allow the P2000's to handle the RAID configuration as i have not found a way to remove the layer
of abstraction for the physical disks through the P2000's to the scale out file servers). Is this achievable through Failover Cluster manager and if it is, in the event of a CSV failure will the failover to the replication CSV be automatic or manual. Lastly
if this is not achievable through failover cluster manager can it be coupled with another native feature like Hyper-V replica, or Storage replica (though it seems this is only available in server 2016). Any guidance or recommendations would be great,
I will continue digging to determine if i can convert the P2000's is to SAS connected JBOD's (If i am able to do this can storage spaces support the functionality i am looking for?). I would be happy to provide any follow up information as this is a rather
complex scenario. Have a lot of clients using storage spaces with Scale out files servers but i haven't yet gone beyond directly connecting a single SAS JBOD. In this scenario i can't really recommend they ditch their SAN's and buy JBOD's.
↧