Etherchannel mystery

SOL10 · ‎03-17-2009

hi guys

i have 2 switches connected in a lab (GNS3) and they are trunked as follows:

SW1 f0/0 -> SW2 f0/0

SW1 f0/1 -> SW2 f0/1

SW1 f0/2 -> SW2 f0/2

SW1 f0/3 -> SW2 f0/3

SW1 f0/4 PC1 - 192.168.50.2/24

SW2 f0/4 PC2 - 192.168.50.3/24

on SW1 i have created 2 etherchannel groups:

Group1 - port f0/0 - f0/2 bundled mode on

Group2 - port f0/3

All ports on SW1 are in trunking mode unconditionally with 802.1q encaps.

On SW2, ports f0/0 - 4 are in trunk mode unconditionally and encaps 802.1q. BUT there is no etherchannel on SW2.

Now i was under the impression the pings from PC1 to PC2 should not be successful as there is no ehterchannel created on SW2 although the ports are in trunk mode on both. (for an etherchannel to work, it should be configured on both Switches)

1) Am i right in my understanding

2) if the answer to the above is i am right, then why are the pings going through and why are they successful ??

am i missing something here??

Thanks in advance

b.julin · ‎03-17-2009

Unless you are running LACP ("active" mode) a single port and a channel don't look much different. SW1 thinks it has links up in the channel group. Probably on group 2, but it could be on group 1 and you were just lucky the hashes worked out and you didn't hit a blocking spantree port.

For it to work CORRECTLY you'll want to match the trunks up. What you're seeing now is a broken setup that just happens to be broken in a way that passes traffic.

Edison Ortiz · ‎03-17-2009

Welcome to the world of Dynamips where all switches' characteristics do not work as expected. You've found one of them.

You won't get that behavior on real switches.

__

Edison.