[c-nsp] Cisco to Juniper LAG/LACP - Juniper not sending traffic over both links.

[Joseph Jackson]

Just got word from the provider.  They are using the default hashing
algorithm which is based on mac address.

On Thu, Sep 6, 2012 at 4:33 PM, Andrew Miehs <andrew at 2sheds.de> wrote:
> Which hashing algorithm are the using? One based on MAC address will do exactly as you are describing.
>
> Andrew
>
> Sent from a mobile device
>
> On 07/09/2012, at 6:58, Joseph Jackson <recourse at gmail.com> wrote:
>
>> Hey networkers,
>>
>> I have 3 sites connected to a service providers MPLS network.  Two of
>> those sites (Dallas and LA) are configured as etherchannel/LACP ports
>> with 2 1gig uplinks in them.  On our end we have Catalyst 6506-E
>> switches the provider side is Juniper gear (I don't have model info).
>> The issue I am seeing is that on our side we are sending traffic out
>> of both interfaces and appear to be load balancing correctly.  The
>> provider (juniper) side is only sending traffic inbound to us over 1
>> of the links and not using the full etherchannel to load balance in
>> coming bandwidth.  On the LA end the juniper routers are seeing two
>> mac addresses learned on that etherchannel - one mac is the SVI and
>> one mac is one of the member ports.  The Dallas side is only seeing
>> one mac address.  The service provider is telling me that their
>> hardware is working correctly and that the hashing algorithm on the
>> juniper boxes is deciding that only one link should receive traffic.
>> Anyone else see this issue or can provide any tips.
>>
>> Thanks!
>>
>> Joseph
>> _______________________________________________
>> cisco-nsp mailing list  cisco-nsp at puck.nether.net
>> https://puck.nether.net/mailman/listinfo/cisco-nsp
>> archive at http://puck.nether.net/pipermail/cisco-nsp/