After upgrading my ESX servers to ESX4 and VC to vSphere4 I discovered the performace of my VCB system had degraded from a virtual backup time of 1 hour to about 4 hours.
This scenario affects those that host VCB on a virtual system and backup the files to a NFS.
After hours of searching and discovering the issue is a known problem by vmware, I discovered a resolve to share here that may help couple of other folks out.
http://kb.vmware.com/selfservice/micros … Id=1012159
The solution I’ve learned from the forums involves utilizing your Sans datastores and adding them to the virtual hardware. If you have a similar setup as mine you may run into a VMFS restriction that will limit your store size to 256GB.
A solution to this will be to create multiple hdd additions then use the extend volume option within the virtual OS/Windows Server to make the drive larger.
I also added a copy command on my vcb batch file that will copy all the backup images once VCB is completed to my original NFS location. To make things a little more automated, I’ve added an RD command to delete the files and folders at the beginning of the batch file so that I don’t run into space issues on the drive.
***************************************************************************************
This was resolved in update2.. It’s actually even faster than it was in 3.5
