HPE Community read-only access December 15, 2018
This is a maintenance upgrade. You will be able to read articles and posts, but not post or reply.
Hours:
Dec 15, 4:00 am to 10:00 am UTC
Dec 14, 10:00 pm CST to Dec 15, 4:00 am CST
Dec 14, 8:00 pm PST to Dec 15, 2:00 am PST
Comware Based
cancel
Showing results for 
Search instead for 
Did you mean: 

Probem with Teaming

 
muralla
Advisor

Probem with Teaming

Hi all ¡¡


I'm trying to install a teaming with a 5500g fabric.


In the 5500g i have lacp enabled in all the ports and spanning tree(rapid).



If i have the 2 NICs  conected to the  fabric the ping does'nt works.


If i have only one of the 2 NICs in one of the unit of the fabric


the ping works...



Anyone can explain me how to conect a teaming with a fabric please?


Its a broadcom teaming.



Thanks all and sorry to my bad english.



4 REPLIES
anacia
Occasional Contributor

Re: Probem with Teaming

For use Teaming in the NIC`s server and connect to switch in fabric, you do not need enable in the port LACP, only connect in both port.  I use LACP when connect between switchs only.



muralla
Advisor

Re: Probem with Teaming

And it works fine?? Spanning tree not cut one of the conections?



 



anacia
Occasional Contributor

Re: Probem with Teaming

Of course, the teaming work well. Spanning tree is more used  for Links between Switchs, when you have several paths.  The control of trafic is in the NIC, if you setting is Load Balancing in the Teaming, both ports in the NIC TX/RX date through the switch.



Fred_Mancen_1
Super Advisor

Re: Probem with Teaming

The correct method to implement teaming (NIC aggregation) using a 5500G-EI switch is creating a static link-aggregation group and assign it to the ports where the server NICs are connected:



link-aggregation group 1 mode static



link-aggregation group 1description SERVER_TEAMING_XYZ



interface giga 1/0/X



port link-aggregation group 1



interface giga 1/0/Y



port link-aggregation group 1



Remember that the ports need to be configured exactly in the same way, only the interface description can be different.



In a scenario of teaming (NIC aggregation with bandwidth increase) If you just connect the NICs to the switch you could force the switch to start a re-learning in the MAC table, what causes high utilization of CPU and memory resources. This occus because the teaming announces a virtual MAC and IP address, but the switch still learns the real MAC addresses from the NICs, binding these real MAC addresses to the virtual IP also, what it means to the switch that there's duplicate MAC addresses with the same IP.



This scenario is different when using failover in the server, when doesn't make difference if the ports are with a link-aggregation group or not.



HTH