Have a UCS B series deployment. Nothing special just dual 6248 FIs. Brocade ethernet switch on the uplinks not Nexus. Because of the limited 10gb ports on the Brocade we want to attach the VNX using appliance ports to the FIs. All designs I found point to 2 seperate VLANs, one in each FI for iSCSI traffic. Plug 1 port from each processor on the VNX into each FI thus giving each processor access to each VLAN.
In order to have full redundancy in case an FI is down, we have created 2 vNICs on the blades and in turn attached 2 iSCSI vNICs to those adapters.
Each iSCSI vNIC requires a unique IQN. so they are suffixed vmware:0 and vmware:1.
This setup works great and provides 4 paths to the boot lun for iSCSI. The problem occurs when VMWare starts. 2 new iinitiators show up in the VNX. This is because in VMWare even though you can split the vNICs into separate vSwitches and VMKernal ports with differnet subnet IPs, they both use a common software iSCSI HBA in VMWare that only uses the IQN vmware:0. Still again the MPIO is working and all 4 paths are up, but now in VNX you have a warning that 2 initiators are down.
As you can see. Annoying, yes. Still functional, yes. My OCD, out of control! :)
It shouldn't be like this!!
