[c-nsp] arps and "it's our address" ?

[Tuc at T-B-O-H.NET]

> 
> Hi,
> 
> On Mon, 2008-06-09 at 11:51 -0400, Tuc at T-B-O-H.NET wrote:
> <cut>
> > He did a "debug arp" and found this :
> > 
> > 002005: .Jun  9 10:44:12.348 EDT: IP ARP: creating incomplete entry for
> >  IP address: 192.168.25.2 interface GigabitEthernet0/1
> > 002006: .Jun  9 10:44:12.348 EDT: IP ARP: sent req src 192.168.25.1
> >   0016.9d4e.9861, dst 192.168.25.2 0000.0000.0000  GigabitEthernet0/1
> > 002007: .Jun  9 10:44:12.348 EDT: IP ARP req filtered src 192.168.25.1
> >  0016.9d4e.9861, dst 192.168.25.2 0000.0000.0000 it's our address
> <cut>
> > He wanted me to post this and ask if anyone has an idea why its saying
> > that. Googles lead him to believe there is a MAC address duplication
> > somewhere.
> 
> It looks more like a loop of some kind. Does spanning tree say anything?
> I don't think it's a duplicate MAC, since it exactly the ARP request
> that comes back.
> 
	Sorry, I forgot to follow up on this. A few hours later he
says that he changed the channel on a radio and it fixed not being
able to ping the management IP. (Even though the device directly
plugged via CAT-5 into his router and wouldn't ping, and plugged via 
CAT-5 to his laptop and wouldn't ping.)

	I wouldn't doubt it could have looped. When he first talked
about the setup, I remember he said he had a switch that he plugged
everything into. As for "everything", it was both Gig ports, his
personal computer, a personal AP, the AP for "Site A" that was
intended for 1 of the Gig ports, and the backhaul that brought 
"Site B" and "SiteC" to the other Gig port. Not sure he ever
changed it. 

	But, apparently, changing the channel of the radio got
the arp working between the device , a CAT-5, and his router.

	Sorry folks... I was asked to keep out of debugging
it, just to pass it along. 

		Tuc