Has anyone else experienced this problem with the "Intel 82574L" onboard network interface on Supermicro X9SCM-F-B?
Basically what's happening is that the network interface will lock up. The machine stops responding, when logging in through the dedicated IPMI or console the network interface shows 99% usage in task manager but does not have a connection to the internet. It doesn't show a red disconnect X it just say's no internet access. If we try to run diagnostics it fails. If we try to disable the network interface it takes a long time, sometimes it disables, sometimes it doesn't. The only way we have found to fix this problem is by power cycling the machine. If we try to restart the machine properly it just hangs at shutting down. However the next day or some other random time the network interface will lock up again.
We recently deployed 11 of these machines with identical specs, 9 of them have Windows Web Server 2008 R2 64-bit and 2 of them have CentOS 6x 64-bit. We've primarily seen this problem occurring on 2-3 specific machines with Windows but it has happened on several other machines less often as well. It has occurred on one of the CentOS machines 2-3 times but not recently. The machines all had Bios 1.0c when deployed but we have since updated to 1.1a which made no difference.
Machine specs:
Supermicro X9SCM-F-B Server Motherboard
Intel Xeon Quad-Core E3-1230 3.2GHz 5GT/s 1155pin 8MB CPU
Kingston KVR1333D3E9SK2/8G DDR3-1333 8GB (2x4GB) ECC CL9 Memory Kit
2 x Western Digital Caviar Black WD5002AALX 500GB SATA3 7200rpm 32MB Hard Drive
BIOS Revision: 1.1a
Intel PROSet Version: 16.6.126.0
Driver Version: 11.11.43.0
Latest INF Drivers installed
All Windows Updates installed
We originally had a Dell Powerconnect 6248 switch but it was quite old so we replaced it with a brand new Dell Powerconnect 5548 switch with latest firmware. The problem occurred with both switches so we don't think it's that unless it has something to do with auto-negotiating or the power saving features.
Things we are trying now:
- Switched one system to use the secondary nic "Intel 82574LM", so far hasn't locked up in 3 days (good sign).
- Rolled back to the Microsoft default drivers, will have to wait and see what happens.
- Disabled Flow Control and EEE (Energy Savings) on the switch.
Is there any way to run a debug or setup network interface logging to find out whats causing this to happen?
Thanks for reading.