[ewg] ib-bonding release 23

[Moni Shoua]

Scott Weitzenkamp (sweitzen) wrote:
> Can you elaborate on "improve failover mechanism", will it improve the 
> problems I see in https://bugs.openfabrics.org/show_bug.cgi?id=541?
>  
> 
Hi,
The improvements is in sending grat. ARP 3 times in 100ms gaps.
This improves the chances that grat. ARP will make on their way to the 
fabric.
Although I intend to keep this workaround in the code (it only can help) 
I don't want to make it a solution for the problem. Please look for a 
post I just sent to the list (with subject "A question regarding 
multicast send") to see what the problem is.

I can't really tell if this improvement will solve the problems in 541. 
What I see in bonding + OFED-1.3 is not seen with previous OFED.
One more word about the test. When iperf reports an interval of 105 
seconds it doesn't make sense to me. I would use a simple ping to see if 
port failover is successful.