[c-nsp] etherchannel problems

[Christopher E. Brown]

I does seem like dynamic would be more risky, but in practice I have
found that running LACP is alot better than "channel mode on".  It takes
a few seconds longer to start up, but does a very good job of protecting
against unbound interfaces.


Holemans Wim wrote:
> We just got bitten by a serious etherchannel problem : we have an 2 gig
> etherchannel link between 2 campus.
> Someone on the other end misconfigured an interface (typed 6/1 instead
> of 1/6)  and had overwritten the allowed vlans on one of the interfaces.
> As a result of this, the interface was thrown out of the bundle (at that
> side only) BUT the interface stayed UP. On the other campus, both
> interfaces 
> stayed in the bundle with very big problems as a result : the 6500 at
> that side considered both lines as valid and distributed the packets
> over both interfaces, sending half of the traffic in 'space'. 
> 
> If the interface had gone down as a result of the unbundling, there
> would have been no problem. We only use static channel settings, so not
> etherchannel negotiations between switches. Can this be solved with
> dynamic etherchannel bundling ? Or someone has another solution for this
> problem ?
> 
> Wim Holemans
> Networkservices University of Antwerp
> _______________________________________________
> 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/


-- 
------------------------------------------------------------------------
Christopher E. Brown   <chris.brown at acsalaska.net>   desk (907) 550-8393
                                                     cell (907) 632-8492
IP Engineer - ACS
------------------------------------------------------------------------