Working with a customer recently I got to spend some quality time troubleshooting VMware Consolidated Backup framework. Generally VCB is a very straight forward install and it pretty much “just works” – which made my recent experience very atypical (in my experience anyway).
Here’s the setup – we have a group of ESX servers, and a Windows Server 2008 Standard 64-bit system, all attached to the same Fibrechannel SAN, with everything zoned properly. VCB is installed on the 2K8 system. The 2K8 OS sees all of of the VMFS luns which are presented to it. We are using Win2K8’s native MPIO stack.
Running VCB Mounter in SAN mode returns an error that there is no path to the device where the VM is stored. Running it in NBD mode works great…except that it passes all of the traffic over the network which is not desirable.
Again, diskpart, and the disk Management MMC see all of the LUNs with no issues.
VCB’s vcbSanDbg.exe utility however see no storage. None at all.
We tried various options – newer and older versions of the VCB framework (btw, only the latest 1.5 U1 version of VCB is supported on Win2K8). We tried various ways of presenting the storage. We even tried presenting up some iSCSI storage thinking maybe it was an issue with the systems’ HBA’s.
Ok, if you’ve read the subject of this post then you already know the answer. In case you didn’t here it is – the system has User Access Control (UAC) enabled. The user we’re running the framework as is a local administrator on the proxy, but that’s not enough to allow it to properly enumerate the disk devices. In order for the VCB framework to work you either have to run it in a command window with the “run as administrator” option, or turn off UAC on the server. The former can be a little tricky to accomplish if you’re wanting to run the framework from inside a backup application, while the latter seems to be the most common approach.
That’s it. Turn off UAC and reboot the computer. Now VCB works great.