Solved: Support Redundant NIC cards in a HP server

steliossideris · ‎09-15-2006

Hi

I would like to connect a HP server with two NIC cards (for redundancy) to two CISCO 2970 switches.

From the HP site they configured Channel Bonding

[root@sm36s_donald root]# ifconfig -a

bond0 Link encap:Ethernet HWaddr 00:14:38:EE:6C:6F

inet addr:172.23.160.36 Bcast:172.23.160.63 Mask:255.255.255.192

UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1

RX packets:563837 errors:0 dropped:0 overruns:0 frame:0

TX packets:110645 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:0

RX bytes:49833329 (47.5 Mb) TX bytes:17943723 (17.1 Mb)

eth0 Link encap:Ethernet HWaddr 00:14:38:EE:6C:6F

inet addr:172.23.160.36 Bcast:172.23.160.63 Mask:255.255.255.192

UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1

RX packets:563837 errors:0 dropped:0 overruns:0 frame:0

TX packets:110645 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:1000

RX bytes:49833329 (47.5 Mb) TX bytes:17943723 (17.1 Mb)

Interrupt:25

eth1 Link encap:Ethernet HWaddr 00:14:38:EE:6C:6F

inet addr:172.23.160.36 Bcast:172.23.160.63 Mask:255.255.255.192

UP BROADCAST SLAVE MULTICAST MTU:1500 Metric:1

RX packets:0 errors:0 dropped:0 overruns:0 frame:0

TX packets:0 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:1000

RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)

Interrupt:26

lo Link encap:Local Loopback

inet addr:127.0.0.1 Mask:255.0.0.0

UP LOOPBACK RUNNING MTU:16436 Metric:1

RX packets:8239071 errors:0 dropped:0 overruns:0 frame:0

TX packets:8239071 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:0

RX bytes:670754425 (639.6 Mb) TX bytes:670754425 (639.6 Mb)

What I have to configure from the Cisco switches site in order to avoid ARP problems?

Thank you for your support

Stelios

anar · ‎09-15-2006

Hi,

Do you have any stack port at the rear of these switces? If yes, you can just connecte them by stack cable, and will get one physically separated switch. There will not be any arp problem. If not , you have to configure bond interface to not load balance, but to be redundant. In this case one adapter for example 1Gbit/s forced by the driver to be 100 Mb/s and second one remains 1Gb/s. In case of link fault, the second adapter will be switched to 1Gb/s mode.

Hope, this will help.

View solution in original post

anar · ‎09-15-2006

Hi,

Do you have any stack port at the rear of these switces? If yes, you can just connecte them by stack cable, and will get one physically separated switch. There will not be any arp problem. If not , you have to configure bond interface to not load balance, but to be redundant. In this case one adapter for example 1Gbit/s forced by the driver to be 100 Mb/s and second one remains 1Gb/s. In case of link fault, the second adapter will be switched to 1Gb/s mode.

Hope, this will help.

Edison Ortiz · ‎09-15-2006

Stelios,

You need to configure Etherchannel. Please read this URL for more info

http://www.cisco.com/en/US/tech/tk389/tk213/tsd_technology_support_protocol_home.html

____

Please rate helpful posts.

Thanks

sachinraja · ‎09-15-2006

Stelios,

We have had issues when trying to load balance between the NICs. when in active/active, the virtual MAC reflects in both the connected switch ports, and the packets tend to loop between the ports. There will be a lot of Spanning Tree TCNs generated, which can shoot up the CPU. We had a 4507R with SUP IV, which got loaded to 70% when doing load balancing. when we made it active/standby on the NIC, the CPU came down to 1 %...

so just be careful on this. best way of doing it is active/standby

hope this helps.. all the best.. rate replies if found useful..

Raj

steliossideris · ‎09-18-2006

Hi Raj,

Thank you for the info.

I have asked from my HP experts to configure the NIC cards in active/standby modeand everything seems to work fine now.

Thank you for your support

Stelios