recently I had to build an environment with Hyper-V virtual fibre channel. As I’m not a SAN expert I had some key learnings that I’d like to share with you.
A good lecture is http://social.technet.microsoft.com/wiki/contents/articles/18698.hyper-v-virtual-fibre-channel-troubleshooting-guide.aspx
So I used a 2 node cluster with one Emulex LP adapter in each node. HA enviroments usually have at least 2 or more.
Firs,t use Hyper-V Manager, Virtual SAN Manager, to Create a New virtual Fibre Chanel Switch
Do that on all cluster nodes with identical Name. Don’t forget to make the check box to link this virtual switch to your HBA.
Second, configure the physical HBA to use NPIV. The Tool for Emulex is One Command Manger (OCManager)
Do this on all nodes. Also ensure to have current Firmware up to date. For Troubleshooting check if Adapter is capable to be used (gwmi -n root\virtualization\v2 Msvm_ExternalFcPort -Property InstanceId, WWPN, IsHyperVCapable)
Third, add a Fibre Channel Adapter to the VM, and configure Word Wide Port Names.
Of importance is the World Wide Port Name (WWPN). This is the End to End connection for the SAN LUNs and must be made known to your switches (zones) and in your Storage to map the LUN.
We use 2 WWPNs to allow for faster migration, so the second WWPN can be made active on the destination while the migration is ongoing. This WWPN is reported to the HBA so that he actually represents the virtual port in addition to his physical port. Failure to open this port on the HBA result in a failure to tart the VM like this:
'vRDGW' Synthetic FibreChannel Port: Failed to start reserving resources with Error 'Insufficient system resources exist to complete the requested service.'(0x800705AA).
'vRDGW': Operation for virtual port (C003FFEFACDC0006) failed with an error: No physical port available to satisfy the request (Virtual machine ID D7AAA601-D3C7-4007-AC9D-C0AA1B087CC5).
This failure was resolved for me with actual FW on the HBA. You also see this failure when the HBA is not connected to a NPIV Switch.
Hyper-V suggests a WWPN out of a pool that you can modify with Virtual SAN Manager. This has to be unique in your SAN
Hyper-V also suggests a WWNN which is always C003FF0000FFFF00. This does not seem to be an issue, however you may see complaints from Cluster Verifier. So use different WWNN for each VM.
During startup of the VM LUNs are detected. If no LUN is available this might be because the WWPN is not correctly in your Zoning, or not mapped on the Storage. This causes the startup to hang 90 seconds at 1%
'vRDGW': No LUNs have appeared for Synthetic Fibre Channel HBA Fibre Channel Adapter (F5F8DD0B-5EF9-4958-9AF3-48FBF3A6C505). The VM has been started after the timeout period (90 seconds).
Now, you should see the LUNs appear in your VM, that needs to be a Server OS. You may see the LUN multiple times due to multiple pathes in your SAN. So usually you will need to install the MPIO and DSM of our storage vendor.
I hope that may help you to get started with vFC.