M1000e / M630 blades and M6220 Swithes

Hey there,

I recently picked up an M1000e chassis with dual M6220 modules in slots A1/A2 and dual MXL modules in slots B1/B2 along with 4x M630 blades - identically configured.

As I understand it, the blades are auto assigned to ports on the M6220's through Fabric A. Thus, I installed ESXi on each of the four blades and ALL 4 blades show connectivity on vmnic0/1 (which confirming via MAC address are the adapters going to Fabric A and the A1/A2 modules.

My problem begins here. The blades in Slots 1 & 2 - call them "server 1" and "server 2" show as having "connnected" network adapters in ESXi, but on the M6220, Ports 1 & 2 are in a link state of DOWN. The blades in slots 3 & 4 ("server 3" and "server 4") also have ESXi showing "connected" network connections and the M6220 reports Ports 3 & 4 as link state "active" - 1Gbps links. 

I have tried moving "server 1" to Slot 5 on the blade chassis - Port 5 on the M6220 remains in a down state. I have moved "server 3" to Slot 1 on the blade chassis and then Port 1 on the M6220 goes into an active state. I also have tried removing "server 4" completely and running with "server 1" in Slot 4 and only running 3 blades - then I only see Port 3 as having an active link state on the M6220 (*note: when I say the M6220, I am referring to BOTH M6220's in slots A1 and A2). 

To me this suggests an issue with the actual blades "server 1" and "server 2" but how do I go about troubleshooting this issue further? All 4 blades have been updated to the latest firmware versions and thus all firmware, BIOS, iDRAC, NIC's, etc. are all running at the same version. 

Does anyone have any insights or thoughts as to what might be going on?? Given that the blade reports an active connection on vmnic0 and vmnic1, to me it suggests there's a problem in the Fabric with the assignment of these blades to the M6220 ports.