Drobo

Windows 2k8 server Hyper V & ISCSI

I am new to the whole vm thing and have been told that Drobo B1200i are the way to go for a budget VM solution using Hyper V.

I am trying to set up the device so that the VMs can reside on an ISCSI share and physical server 1 is pointing at this ISCSI location to run the VM’s from. We would also like to have a 2nd physical server that can point to the same ISCSI target and run the VMS from there in the event that the 1st server goes down for Maint or unexpectedly.

I know there is a great document on how to setup a Windows 2k8 clustered environment for HA, but my manager does not want to go that route. I know without the environment having a shared Clustered store and server resources, having 2 windows servers talk to the same ISCSI target would be a disaster as there is no control over reads and write to the target disk and it will corrupt the disk.

What is the best way to utilize the Drobo and have manual “failover” in my situation? Im not sure how this would or could work.

thanks

I know there is a great document on how to setup a Windows 2k8 clustered environment for HA, but my manager does not want to go that route. I know without the environment having a shared Clustered store and server resources, having 2 windows servers talk to the same ISCSI target would be a disaster as there is no control over reads and write to the target disk and it will corrupt the disk.

Yes, you absolutely need to enabling clustering for this usage case. The good thing is that HyperV makes setting up a cluster easier, and includes a certification test to ensure the cluster is setup correctly before allowing it to be used.

B1200i supports clustering with HyperV

Jason

Thanks Jason. My other issue in that case is that we are running Windows 2K8 R2 Standard, and the failover clustering feature is not available in this edition of windows server. guess im out of luck as there is no licensing $$$ in the budget

If you can’t set up an active cluster, you’ll just need to manage the failover by hand.

By default, a LUN only allows a single host to connect to it. If you know your primary host will fail in a way where it will go offline (because you switch it off for example), then the iSCSI connection will timeout eventually, and you’ll be able to attach to the LUN from the backup system manually, and remount the LUN and volume.

Clearly there might be other system level issues you’d need to address in that case, such as telling your users how to mount shares on a different server, which a clustering solution would take care of.

Its probably worth doing some failover experiments to understand the workflow before the big day!

Cheers!

Jason