[c-nsp] etherchannel load-balancing & WS-X6708 issue

[Jiri Prochazka]

>-----Original Message-----
>From: Nick Hilliard [mailto:nick at foobar.org] 
>Sent: Monday, September 12, 2011 11:13 AM
>To: Jiri Prochazka
>Cc: cisco-nsp at puck.nether.net
>Subject: Re: [c-nsp] etherchannel load-balancing & WS-X6708 issue
>
>On 09/09/2011 12:33, Jiri Prochazka wrote:
>> Primary box is 6500-1 (several hunderd Vlans and SVI's) and under 
>> standard circumstances Po1 is the only path utilized, but even if I 
>> freak out, I am not able to push more than aproximately 25 Gbps from 
>> 6500 over Po3 to edge-2. My initial guess was bad port assignment and 
>> therefore not utilized local switching..
>
>what is the output of: "show etherchannel load-balance"?  Also, is the port
channel between 6500-1 and 7600-1 configured as a switch port or as a routed
port?  This makes a very substantial difference in how traffic is load
balanced over the bearer links.

Port-channel is switched, load-balancing scheme is src XOR dst IP.

>> The major problem which I can not beat (and maybe it's a dead end for 
>> me) is the system how IOS decides which ports will be assigned to 0-7 
>> ID's used in etherchannel load balancing algorithm. As I observed, it 
>> absolutely depends on a sequence of adding/removing ports from/to an
etherchannel.

>Have you played around with the "port-channel hash-distribution" command?
>This may help to make the ID assignment more consistent, even if you don't
have full control over everything.


Whoa! This did the trick! When I set hash-distribution to 'fixed', ID's are
assigned always the same. I just had to reorder ports between 6500-1 and
7600-1 and whole traffic is switched localy without hitting the fabric!


Actual egress rate from module 2 is around 30 Gbps, latency is ok and..

7606-1#show platform hardware capacity fabric
Switch Fabric Resources
  Bus utilization: current: 3%, peak was 43% at 13:46:21 CEST Wed Jun 29
2011
  Fabric utilization:     Ingress                    Egress
    Module  Chanl  Speed  rate  peak                 rate  peak
    2       0        20G   14%   48% @10:42 11Aug11   34%   65% @10:44
11Aug11
    2       1        20G   17%   49% @18:17 11Jul11   36%   69% @19:33
25Aug11



Thank you, Nick!


>Sounds like you may just have too much traffic for your hardware
configuration to make it work reliably.

>Nick