Server Products
Data Center Products including boards, integrated systems, Intel® Xeon® Processors, RAID Storage; and Intel® Xeon® Processors
Announcements
Welcome to the Intel Community. If you get an answer you like, please mark it as an Accepted Solution to help others. Thank you!
4302 Discussions

Teaming Intel PRO/1000 PT Dual Port PCI NICs in Windows 2008 Standard 64-bit

idata
Community Manager
979 Views

I support a Sun Blade X6270 server, using two http://www.oracle.com/us/products/servers-storage/storage/storage-networking/042897.html Sun StorageTek SG-XPCIEFCGBE-E8-N Dual 8 Gb FC Dual 1 GbE HBA ExpressModules (in fact these are http://www.emulex.com/products/lightpulse-hbas/sun-branded/sg-xpciefcgbe-e8-z/management.html Emulex LPEM12002E-S HBA/Ethernet cards repackaged and sold by Oracle/Sun. The operating system is Windows 2008 Standard 64-bit with Service Pack 2.

Windows reports 1 GB ethernet ports on these cards as Intel PRO/1000 PT Dual Port NICs. They are currently using the latest available Intel drivers (version 9.13.41.0 dated 2010-03-26). PROSet version 16.2.49.0 is part of the installed driver. A single port on each PCI card is connected to a port on Cisco switch.

I am trying to team together one Ethernet port from each PCI card for adapter fault tolerance. I can create the team without any problem, but once it is created and IP address is assigned to it (no VLAN), then an interesting problem occurs. I can ping all hosts on the same subnet with the last IP octed being 1-120, but when I try to ping any host above 120, the connection fails. Network support staff informes me that they see a lot of packet being dropped, when the team is active.

The moment I disable one of the team members, even without removing the team, everything works fine and I can ping full range of host IP's on the subnet.

I would appreciate any help in resolving this problem and providing full functionality to the NIC team.

Regards,

Romuald Czajkowski

0 Kudos
1 Reply
Daniel_O_Intel
Employee
72 Views

If you leave the members active, but pull one of the two cables - can you ping everyone within the range again?

The fact that pings above 120 fail might be a red herring, or it might be one bit off in a subnet mask - you aren't using 255.255.255.128 anywhere, are you?

What do you see on a 'debug all' on the cisco, when this happens?

Reply