Synthetic FibreChannel Port Failed to finish reserving resources
So its a Saturday afternoon playing my 2 years old and a phone call came to end it ;-), so one of our customer had an issue with live migrating virtual machines with vFC disks attached to it.
The first error I logged was:
Live migration of ‘Virtual Machine XXX-XXX-XXX’ failed.
Virtual machine migration operation for XXX-XXX-XXX’ failed at migration destination ‘XXX-XXX-XXX’. (Virtual machine ID 113F69E7-5591-492E-BD28-C7DDEF2ABB6F)
XXX-XXX-XXX’ Synthetic FibreChannel Port: Failed to finish reserving resources with Error ‘Unspecified error’ (0x80004005). (Virtual machine ID 113F69E7-5591-492E-BD28-C7DDEF2ABB6F)
Digging deeper into the event logs I found out this as well:
‘XXX-XXX-XXX’: Live migration of the machine failed because all the LUNS on the source machine were not seen on the target for virtual HBA Fibre Channel Adapter (8F399E51-2CE5-4572-B814-A5780F0F4FF0). Expected: 6 LUNs, Found: 2 LUNs. Please review the LUN mappings for the World Wide name pair of the virtual HBA. (Virtual machine ID 113F69E7-5591-492E-BD28-C7DDEF2ABB6F)
This made me speculate about the zoning on the FC switches so I went back and reviewed everything and nothing seemed to be out of order, surely I couldn’t do further testing on production virtual machines so I decided to create one of my own:
- Created the VM.
- Added two vFC adapters to it.
- Installed Windows Server 2012 R2 on it.
- Logged the WWPNs for SET A and SET B.
Now on the FC switch I could see SET A addresses, what I needed to complete my zoning was to live migrate this VM so that to get the vHBA to change the WWPNs to the SET B addresses and there was the surprise, even when the VM was not zoned I couldn’t live migrate it and I would still get the same error.
This narrowed it down to the level of the host, rather than being an issue on the FC switches or on the SAN storage itself, so the logical thing to do was updating the drivers/firmware of the HBA.
Taking it further since the hosts were HP DL380 servers I decided to install the new SPP (which is currently 092014) and after installing it on all the hosts (which included the QLogic HBAs updates as well) I was able to live migrate these virtual machines once more.
Another thing which occurred to me is that since the hosts were Windows Server 2012 Standard and the virtual machines were Windows Server 2012 R2 this might cause an issue at some point (its supported here though).
Anyway, I hope this help anyone who’s been facing such an issue.