We're looking to implement NW 9.1.1 in a highly virtualized (VMware) SQL environment and I would like to move towards image-based backups. I realize that not everything should be snapped and we will identify the servers that have high transaction rates, but one of my main concerns is obviously impacting database performance.

In the past, I used a different backup product and was able to filter out all datastores that contained the .ldf, .mdf, and .ndf files, leaving only the C: drive and maybe a utility drive or two to be backed up. I then had a SQL agent installed to backup the database and logs separately. This worked out great because I could now avoid database stuns during the snapshot operation and not have to backup all of the data essentially twice in a night.

So, is there a way to filter out specific datastores, vmdk's, or hard disks using directives? Or am I going to have to go into the Group and drill down into every VM then manually deselect drives that I do not want snapped?