In my server lab, I am running a vCenter cluster with 2 ISCSI targets. Both of these are directly connected to switch using SFP+ port.
I assume, when I do such a process as cloning a VM from one datastore to another (say ISCSI 1 -> ISCSI 2), the traffic would flow ISCSI 1 directly to ISCSI 2 ( ISCSI 1 SFP+ -> switch -> ISCSI 2 SFP+). I think the above is how it works based on common sense, however, if I am wrong please let me know.
I would like to increase the bandwidth between both ISCSI instances. However, the switch only supports 2 SFP+ ports, while each ISCSI has 2 available SFP+ ports(4 total). My question is, is it possible to perhaps increase bandwidth between 2 ISCSI targets by directly plugging them into each other on their secondary SFP+ cables while keeping the initial SFP+ cables where they are at(so the vHosts can actually reach ISCSI targets)? That way VMs could to be transferred over both SFP+ ports:
1) ISCSI 1 SFP+ -> switch -> ISCSI 2 SFP+ 2) ISCSI 1 SFP+ -> ISCSI 2 SFP+
I am not too sure if ISCSI can support such a thing, or if networking, in general, can support this, so please let me know about potential solutions to this problem.