There are several things to look out with this issue.
Firstly, is the server you are trying to connect LUN to a MS Cluster Box? If so do you have SnapDrive installed on all nodes in the cluster? Also SnapDrive when running in a virtual machine should be 6.3 or above, to have correct venter integration.
I assume as you are trying to do mapping from directly inside the Virtual Machine you are using iSCSI? There are a few known issues with MS iSCSI initiator, ms cluster and virtual machines. It is worth reviewing the MS Cluster document on the NEtApp Webpage. If you are using FC for VM connectivity, does your hardware support NPIV? If so this is a way to present virtual FC Card to virtual machine and map at that level? SnapDrive should then be able to operate normally.
For SnapDrive to connect a LUN it needs to have its mappings removed, as you have done already. Depending on your luck, if you map the LUN using ANY method, SnapDrive will sometimes detect the LUN and allow you to manage it providing other conditions have been meet (snap drive on all cluster nodes).
As far am I am aware SnapDrives does not support connecting/creating LUNs where the LUN has been zoned using NPIV to the VM, there has been nothing in any of the manuals regarding NPIV.
With FC we are using the integration between SnapDrive / VSC to present RDM's into our virtual machines, we also have the same issue of not being able to connect physical RDM's into our VM's on Windows2008 but can create/disconnect then connect to the same LUN using Snapdrive within the VM.
We are also using SnapDrive 6.3.1R1 within the Virtual Machine and connection to both VC and SMVI/VSC worked as expected during install.
Windows 2008 R2 SP1
VIrtual Center 4.0
But we see the following error message is shown from SnapDrive: "A timeout of 120 secs elapsed while waiting for volume arrival notification from the operating system" and the disk disappears from DIsk Management - during the connection process we see the size of the LUN and it shows as a Online Basic (Unallocated) disk but then the timeout occurs and it disappears. I've had the same error message when mounting a snapshot using SnapDrive on a Windows 2008 OS and the following fix worked but we are unable to install this hotfix into our guest OS:
Malcompenn is right, last time I checked SDW 6.3 didn't support NPIV so previous versions would definitely not be supported. There were discussions to possibly add to the roadmap in future releases but don't think its in 6.4 either.
The issue I commented above was a user issue and certainly not product issue - we were trying to connect to a LUN which we hadnt migrated data into, therefore the LUN was without an already existing filesystem. Once we actualy migrated the data into the correct destination LUN it "connected" as expected
Testing all functionality of moving the VM with RDM attached between ESX servers no problems what do ever.
Are you getting the error on connect only? Was the LUN you're connecting to created by Snapdrive? If not, you will need to prepare the LUN so SnapDrive can manage it and procedure is in the IAG.
As a test, try to create a new volume, create a new lun with snapdrive, and see if same error exists. If yes, Please contact NGS and open a ticket as they may be able to assist you if it is a issue with the software or with NPIV support.