Virtual Connect (VC) Config with 2 vnet and behavior of Teaming with TLB

[chuckk281]

Cristoph had a customer  issue:

***************

I got the question from a customer, who is balancing the primary port of the Team between the two VC Modules. Now after a failure of the network switch where one of the SUS is connected he asked how to rebalance it again.

1)      From a VC point of view he should be fine, as smart link should do the work

2)      From the host it gets more difficult. I found no deeper explanation/advise in regards to the NCU setup. With the normal TLB setting the primary port should failover to the other for these servers that had the primary one on the failed vnet side.  It seems that there is no automatic fail back.

The questions  :  As we run into this situation with every  Online firmware update. Is there a way to rebalance the load online? Will the primary port  go back to NIC that was chosen as a primary during creation of the team after a reboot?

**************

Input from Robert:

*************

The ‘HP ProLiant Network Adapter Teaming White Paper’ will shed the best light on this subject.  Unless I have been misunderstanding for a few years, when the failed adapter recovers, the load balancing will automatically resume.  But for the Translation Lookaside Buffer (TLB), the receive traffic will remain on the port that was failed over, unless the team was configured for TLB with fault tolerance and preferential order.  Whether the transmit/receive adapter swaps back is dependent on the type of TLB configured.  For TLB specifically, you can read the description on page 36 of the White Paper for recovery.    

This is a great doc by the way. 

ftp://ftp.compaq.com/pub/products/servers/networking/proliant-teaming-whitepaper-march%2008.pdf

*************

Comments or other help?