Etherchannel

Unanswered Question
Oct 30th, 2007
User Badges:

Hello,


Here is my problem:

Traffic going between (2) Distribution switches [6509E - CAT OS 8.5(8)] within the same ServerFarm rides on a 4GB Etherchannel consisting of (4) 1GB ports.

The ports that are channelled are: 1/9-10, 2/9-10 and the channel ID is 1670.

Ports 1/9-10 and 2/9 have traffic at around 40 - 60 Mbps; while, traffic on port 2/10 is pegged at 1GB and is reporting an abundance of out-discard errors.

This condition has been present for 2 weeks and has been impacting performance.

I know that Etherchannel is setup for session based load-balancing and not per packet, but given the volume of traffic, on a single port and the lack of traffic on the other 3 ports all within the same channel, is there a way to check the health of the channel??

Is there anyway, to determine why traffic is only using port 2/10 and not the other ports within the Etherchannel??


Thanks for your help,

Tony



  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
dominic.caron Tue, 10/30/2007 - 12:06
User Badges:
  • Silver, 250 points or more

The load balancing is based on a hash of information found when the flow is created. You can modify the information the hash is base on with the command


set port channel all distribution {ip | mac| session | ip-vlan-session} [source | destination | both]


You can review this link

http://www.cisco.com/en/US/tech/tk389/tk213/technologies_tech_note09186a0080094714.shtml


Actions

This Discussion