As you may already be aware, Microsoft are supporting multi-site clustering through Windows Server 2008 failover cluster services. Microsoft however are not doing the actual replication / mirroring between the storage boxes located at each site, and they are leaving this task to 3rd parties. I have read quite alot on this and the main products which I could find info about were Steeleye Datakeeper Cluster edition and Doubletake Geo-Clustering. These are both software based solutions.
My question is, do NetApp have something where we can have two boxes in separate sites and failover windows cluster services between these two sites? This is what the two software products mentioned above do, but we would like to offer a solution based on netapp hardware.
Can someone please indicate which options we have as all I have found from the NetApp site was always referring to two Hyper-V nodes connected to the same central storage. We need to have two Hyper-V nodes connected to different storages which take care of the replication and appear as a single same storage to both geographically dispersed nodes.
Has anyone found a good technical report from NetApp on this subject? The MS whitepapers (http://www.microsoft.com/downloads/details.aspx?familyid=75566F16-627D-4DD3-97CB-83909D3C722B&displaylang=en) do not really get into the SAN replication piece and the podcast basically says to follow your SAN providers recommendation. Our situation may be somewhat unique as well since we will be replicating data between our two hosted facilities that are connected via layer 2 switching via 10gig lit fiber thus giving us one logical subnet. We have Cisco UCS hardware with 98% of our environment running vSphere 4.0 as well. If anyone finds a good doc, please let me know....I will do the same. Thanks!
We have not implemented a Geo Cluster with NetApp but we are using SnapMirror and Symbolic Links allowing us to link the same VM configuration on two different Hyper-V clusters in different location.
We use BCG to route our production LAN to our DR sight and NetApp snapmirror to replicate our VM and data. Works pretty good and takes about two hours to failover and bring up Exchange, SQL, and Sharepoint on the same network without any modfication to actual VM network configuration.
However, the setup of the Symbolic links is complicated and difficult to do during the DR and requires some up front time to ensure you can meet failover SLA.
An alternate method is to export VM's and use SCVMM to deploy them on your DR Hyper-V cluster. Maintaining two seperate identical VM configuration with same Network configuration while replicating your LUN and VHD's with NetApp SnapMirror.
Also, Citrix Essentilas 5.5 for Hyper-V is advertised to have the same functionality as VMware SRM and has a plug in for NetApp. I do not have experience with it yet and it does requires SCVMM R2 in both sites.
I have the same requirements. Has anyone found a whitepaper or other technical doc from NetApp on this configuration? We are also leaning toward using VCS, but since this is a production HA environment, it would be good to have confirmation from the storage vendor.