Hi Ramon,

that's right, LLDP-based discovery isn't as fast as many applications would
require. You could definitely base your link status logic on port status
changes, however I don't think there is a discovery module that currently
does it. I think the incentive towards LLDP discovery is that it can
discover failures that wouldn't be caught otherwise (case where a link is
effectively/actually down, but a port status changed has not been sent).

That said, note that you can tweak the LLDP probe period. It is currently
set to "TIMEOUT_CHECK_PERIOD = 5" seconds (in discovery.py), so you can
play with that knob, keeping in mind that the bigger the frequency, the
more the control traffic for discovery.

As for keeping track for hosts locations, take a look at the implemantions
of Authenticator
<https://github.com/noxrepo/nox-classic/wiki/Authenticator>and
netapps/hoststate/hosttracker (unfortunately the API for the latter is
I think less rich in python)

On Mon, May 7, 2012 at 6:24 PM, Ramon Marques <[email protected]> wrote:

> Hi,
> I was trying to use the discovery/topology modules to detect link failures
> in the network. But it takes too long to to detect a link failure(at least
> 10 seconds).
> Is it possible to use the port status messages to detect a link failure,
> instead of LLDP messages? Is there a module that implements this
> functionality?
>
> Another question. How can I know the location of all detected hosts(switch
> and port which each host is connected)?
>
> thanks in advance
>
> regards
>
> --
> Ramon Marques Ramos
>
> Engenheiro de Computação
> Mestrando em Informática -UFES
>
>

Reply via email to