Losing Production Network Connection with Dual Nics on different subnets
Posted on 2009-12-28
I have a Systems Center Operations Manger R2 RMS server running Windows Server 2003 x64. The server is an HP BL460c G1 blade with latest firmware/drivers.
My network configuration is as follows:
2 Internal Nics
NIC 1: Prod network with 172.17.*.* IP
NIC 2: DISABLED in windows
2 Nics from Mezzanine card
NIC 3: GPRS Modem attached via Network with 192.168.2.1 IP
NIC 4: DISABLED in windows
So in total, I have 4 NICs, 2 of which are in use. NIC 3 is attached to a cellular modem to perform paging from SCOM.
The configuration works. My production network works fine and the paging works fine. The issue that I have is that after a few days, my production network just stops functioning. The NIC itself is connected, but I cannot ping the server what-so-ever. If I connect via the iLO (remote interface) on the server, the server will be up and running. As soon as I disable NIC 3 with the modem, my production network instantly comes back!
It seems like the production traffic gets confused and no longer tries to go through NIC1. I can immediatley re-enable NIC3 (modem) and everything works fine again for a few days. NOTHING is logged in the application or system log stating what happened. There will eventually be events stating that the server can't communicate with a DC and you can tell when SCOM stops working, but no events state why the traffic stops communicating via NIC1.
My Binding order is set in the following order:
1. Production Network
3. Disabled NIC
4. Disabled Nic
Note: If I leave the modem NIC disabled, it will function fine and I have not experienced the issue.
Other note: I completely removed the HP teaming software hoping that it might be it, but that didn't help either.
Does anyone have a clue why this might be happening? Please let me know if you need any additional information.