Pro 1000 Issues

System Specs for machine having issues:

Motherboard: H87WS-DL R
CPU: Pentium G3220
Ram: 16GB
Raid Controller: Areca ARC-1210
OS: Server 2008 R2 with all patches installed.

Network setup for multi-path iSCSI
Vlans 200, 201, 202, and 203 for iSCSI setup in my HP Procurve switch.
1 port dedicated to each VLAN to be used for the above server with quad nics
iSCSI Networks: 192.168.200.0/24, 192.168.201.0/24, 192.168.202.0/24, 192.168.203.0/24


Problem: I have the above system setup to be my VMware vCenter Server and Starwind iSCSI server for my VMware lab, I within the last month migrated to this solution from Openfiler after I got tired of it's iSCSI flaws. I have several Intel NICs I have tried to use in this machines, a Dual Port Pro 1000 PT, a Quad Port Pro 1000 PT, and a Quad Port Pro 1000 ET. With both of the quad ports. Initially when I migrated from Openfiler, I had the Pro 1000 PT in the system, so I installed Server 2008, used one of the onboard Realtek nics for management, and configured each of the ports on the Pro 1000 to each of the for iSCSI networks. Everything was working fine, could ping my switch on those VLans, as well as ping my ESXi hosts. I then setup Starwind and when to start adding the storage volume to the ESXi hosts and they couldn't find it. Further investigation found that I could no longer ping anything on the iSCSI networks, including the local IP for those ports, (192.168.200.5, 192.168.201.5, 192.168.202.5 and 192.168.203.5) I then found in the event log an error about (The network driver has been stopped because the network adapter has been removed.) I tried disabling and re-enabling the NICs and they just disappeared. I rebooted and tried again, got a little further, was able to add the volume to the 1st host, but upon trying to add to the second the network dropped again. This time I noticed the network utilization in task manager was at 100%. I wasn't able to find a lot about this type of problem, so I chocked it up to possibly a failing Quad NIC, I swapped it out for a Dual Port Pro 1000 PT, added a PCI NIC for management and configured the 2 intel ports and 2 onboard realtek ports to be iSCSI. This worked fine for a week and the Intel nics dropped out with same symptoms as above. Today I swapped out the intel Pro 1000 PT dual port for a quad port Pro 1000 ET, with in minute of getting the NICs configured they dropped out with all the same exact symptoms, I have even gone as far as trying to disable the onboard NICs and that hasn't helped. I have tried various driver version from Intel, nothing has seemed to help, I have run out of ideas.

Try the network cards in another machine to see if the issue is with the hardware.

In the server, if both cards are dropping at the same time, the cards are likely fine. Have you tried other PCIe devices to see if something on the motherboard is failing? A video card would make it very obvious if something was going wrong.

I hadn't thought of trying that, but what I found odd was I had been running the system fine with Openfiler using the Quad Pro 1000 card for several months, what drove me away was the iSCSI instability that Openfiler has and in order to fix that issue you have to pay for the patch.

Swapped slots between the Raid Card and the Quad NIC, same issue, works for a few minutes, start to put any usage through the ports and they go dead.

Figured it out! It turns out it was "Link State Power Management" I turned that off and I have now been running for two days so far with no NIC dropping out. I ended up digging more into the bios looking for an issue when I upgraded my Areca ARC-1210 to an LSI Sas 9261-8i raid card and the system would randomly just freeze when in windows.

Glad you got the issue figured out, I've never seen that option cause an issue.