Today, I was preparing an environment for a physical relocation. My employer is moving to a new office building, and so the complete IT needs to be moved, too. We’re currently running all services (except a now defunct fax server and a Cisco CallManager) on four Dell PowerEdge 1950’s, attached through Fibre-Channel to a Dell/EMC CX3-10c SAN using Fibre Channel only.

Our regular backupsolution won’t allow for backup or recovery of complete VM’s, so to make sure no data gets lost, I wanted to create backups for disaster (say, the SAN hurls down a case of stairs by accident) recovery (I want my data back!) purposes. I’ve used and loved Veeam Backup and Replication for loads of other customers, so I wanted to give it a try in this scenario.

Installing and configuring the software was a breeze, and within thirty minutes, I was making backups of my VMDK’s. Veeam happily grabbed loads and loads of data, but it didn’t perform as I had hoped:

VCB-NBD

As I have installed Veeam in a virtual machine, I couldn’t leverage VMware VCB’s ability to access the SAN directly and backup using VCB’s SAN mode. Instead, I had to use NBD mode, which is a lot slower. But wait! Wasn’t there a way to attach VM’s to the fabric directly?

With NPIV, you can do just that! I powered the VM off, enabled NPIV, and booted it up again. While that was running, I checked the physical HBA’s, Fibre Channel switches and the CX3-10c for NPIV support. All was well, until I got to the point that the SAN didn’t recognize the newly created WWN. It simply did not show up…

After some Binging Googling, I found out that NPIV only works with RDM’s, not with direct access to the VMFS datastores. Mister Boche has written all about it here.

Bummer, that craps all over my plan to back up data with VCB’s SAN mode. Back(up) to square one…