Hi,
Well I was before working for a company, that also had 2xESXVCVCBproxy (CA Brighstor) installed. we were running about 40 production VMs. We used VCB to back them up (fullvm). File-level backup is stupid, as there can be no differential/incremential backup, as all mounted files are "new" every time snapshot taken :smileyhappy:
We started to create vcb backups using integration modules for CA, but this was f*cked up, CA didn't know how to set -M 1 option to make 1vmdk for each disk, VMware said - ask CA, as this is their product.
Finally I made such solution:
On our backup server, I setup 500GB volume (internal disks), and prepared bat file that did command vcbmounter.exe for every VM needed to backup (I needed to backup only 20VMs). Every VM was backed up with fullvm type and -M 1 mode (1vmdk/disk) and -L 6 (biggest logging).
It run every niight, starting at 11PM, and finishing about 4-5AM.
at the end of this script I put script that sent email about which folders were backed up (vbs that compared folder names created on backup server with VM names that should be backed up). and It also sent logfile created.
So every morning when I came to work, I checked my email, saw that for example VM1, VM5 and VM7 were missing on VCBproxy volume, so this mean that they need to be "checked" why. Then I took a look at log and saw frequent entries about SCSI reservations. So I needed to fix this issue. Then fixed these VMs(often hung snapshot) and repeat this script manually.
I know in big environments it is not an option (too much manual work), but if you have storage like EMC, Netapp or other big solutions, then their SNAPSHOTS are enough to backup this data instead of using VCB