network reconfiguration?

[gstephenson]

I have a two-node hyper-v setup running v5.8 on 2008R2 as follows:

NIC1, port 1: public network (100Mb to switch) for Starwind heartbeat and web traffic
NIC1, port 2: unused

NIC2, port 1: 192.168.10.10 (1Gb crossover) for cluster heartbeat
NIC2, port 2: 192.168.20.10 (10Gb crossover) for Starwind sync
NIC2, port 3: 192.168.30.10 (10Gb crossover) Starwind data traffic (not sure??) and Live Migration
NIC2, port 4: unused

The problem is that every so often NIC2 will spontaneously reset itself. It is an HP NC375i. Google it and you will find loads of problems with this card. Sometimes I have to reboot to get it working again, other times everything recovers on its own after doing a full sync. I'm not sure if I will be allowed to replace it right away, so I want to know if the following setup would work:

NIC1, port 1: public network (100Mb to switch) for Starwind heartbeat, web traffic and cluster heartbeat
NIC1, port 2: 192.168.10.10 (1Gb crossover) for cluster heartbeat, Starwind sync and Live Migration
NIC2 would be disabled.

Traffic is very low and we rarely ever need to do live migrations. I know there isn't any redundancy in either of these senarios but this is all I've got. Would this be valid for both Starwind and hyper-v failover clustering? Would I be safe from split-brain problems? (NIC1 is a HP NC382T)

Thanks...

[Anatoly (staff)]

Dear gstephenson,

Yes, you are 100% right when saying that this configuration is not too good.
I can even say more - 100MBps data link is lower then numbers mentioned in our System Requirements (please see the link below).
http://www.starwindsoftware.com/system-requirements

But still, this configuration should work (it hasn`t been tested in our test lab).
Our very strong recommendation is to use at least 1 Gig NICs.
Also make sure that synchronization channel traffic will be utilized for data synchronization purposes only.