the limit for vmfs and RDM volumes is 2tb - 512 Bytes. There are a couple ways to do what you are accomplishing:
One is to create multiple 1.99TB Volumes and use Extents to create one big volume (which can be as much as 64TB). you format this with 8MB cluster sizes and then create the appropriate sized VMDK on it and migrate your data (although I think you still hit the 2TB limit). Using Robocopy you can keep the permissions and do an incremental copy over a month or something until all data is in sync. By disabling strict name checking you could then have the new server allow users to map using old DNS cname. You could also implement DFS replication as well, although thats usually harder to implement on existing configurations.
Extents are generally discouraged as you can never remove the disk later without wiping the entire vmfs volume. Personally I think of Extents like Microsoft Dynamic disks...sure it can be done, but its probably not a good idea...
Another option (and this I generally recommend as its better for backups) is to tray and split up your data into different volumes so you can have multiple 2 TB volumes. The biggest advantage to this is that when you have backup software that uses multiple streams (like hp data protector or TSM), you can put 5 different 2TB volumes in a single backup job and they will all backup at once, vs a single job only being able to use a single thread against a single volume. Cuts down backups times dramatically...
Obviously this all depends on you file system and share layouts.
If you have the capabilities to turn that other server into an ISCSI target you could also use the ISCSI initiator right in the VM operating system to mount that disk as a raw volume...of course it would be better sitting on a proper SAN..but beggers can't be choosers sometimes 🐵
