On Fri, 11 Jul 2014 14:42:07 -0300 Bruno Galindro da Costa <bruno.galindro@gmail.com> wrote:

>>
>> You are referring to three different addresses here, not two, which is a
>> bit confusing. I'm going to assume that whenever you wrote 250.121,
>> what you really meant was 252.121.
>
>
> Yes! The correct IP adress is 252.121. Sorry :)
>
>> So, my question is: If you search for 00-1E-C1-ED-78-80 in the Machine
>> Tracker, do you have any hits for 251.245? If you look at the complete
>
>> interface report for 251.245, do you see any interfaces with this adress
>> in the physical address column?
>
>
>
> No for booth, and now I know why: 00-1E-C1-ED-78-80 is the chassis MAC

> Address os 251.245 switch. The mac address of 251.245 is 00:1e:c1:ed:78:8*1*.

> This is because 251.245 is the IP address of a VLAN interface... The device
> was registered with category SW not GSW. After change the device
> category from SW to GSW, waiting a few minutes to ipdevpool run, the
> machine tracker could find the relation beetween 251.245 and

> 00:1e:c1:ed:78:8*1.*

I'm not immediately sure why that helped, but I'm currently at
EuroPython in Berlin inbetween my vacation weeks, so my memory on this
issue may be foggy ;)

A bridge MAC address is retrievable via the BRIDGE-MIB, but NAV doesn't
currently use this for anything.  If this address doesn't correspond to
any of the physical interfaces that are collected, and is never used for
IP communication, NAV will not see it. This may be what was the case
here.

It may be that we should collect and store the bridge address somewhere
to aid in identication. It may also be that one of the auxilliary tables
of the LLDP-MIB will provide more information. From the snmpwalk dumps
you sent, it appears that there was more information to be found about
the remote hosts discovered by LLDP.


If you would file a bug report we won't forget about it :)

--
Morten Brekkevold
UNINETT