Quoting Peter Rathlev <[email protected]>:
On Fri, 2011-07-29 at 14:32 +1000, [email protected] wrote:
Have 2 100Mb eth ports in portchan, and 1 port seems to be doing all
the receiving:
Keep in mind that ether-channel load-sharing is a transmit function. If
one switch receives an uneven load the fault lies in the other end. In
your case it's ESW01 (the 2950) that doesn't load-share adequately.
The only thing you can do is to try changing the algo in the 2950 with
the global config command "port-channel load-balance dst-mac". This
might or might not give you better sharing. The 2950 is rather limited,
accepting only dst-mac or src-mac.
Thanks Peter - Yes, the 2950 is limited...it will be replaced in the
next couple of months with a 3560
What kind of traffic traverses this link? Maybe there's another way to
split up the load.
Traffic is a mix of inet/vrf...the main culprit is a backup service
that replicates basically 100% of the time at whatever bandwidth is
allowed...potentially moving there vlan to a dedicated trunk would
help, but id prefer to have the portchan used...if I cant, then ded.
trunk, then go gbit between the switches when we upgrade the 2950 to a
3560.
-------------------------------------------------------------------------
This e-mail was sent via GCOMM WebMail http://www.gcomm.com.au/
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
