[Olsr-users] duplicate hna messages

Henning Rogge (spam-protected)
Wed Sep 8 16:33:27 CEST 2010


Am Dienstag 07 September 2010, 17:22:52 schrieb John Hay:
> On Mon, Sep 06, 2010 at 12:30:44PM +0200, Henning Rogge wrote:
> > On Mon September 6 2010 11:48:19 John Hay wrote:
> > > > This looks like one or a few of your OLSRs don't do the duplicate
> > > > detection correctly.
> > > > 
> > > > How large is your network ?
> > > 
> > > About 30 nodes.
> > 
> > In a network like this you  should never see OLSR TTLs of less than
> > 255-29, unless you use fisheye.
> 
> What will happen if you use fisheye?
OLSR generate TCs with a start-TTL of 2, 8, 16 and 255. But you still should 
not get something like a TTL of ~100 (because your mesh has less than 150 
nodes).

> > > It does not happen the whole time. It seems like sometimes
> > > it would be ok for more than a month and then something triggers it.
> > 
> > Some of the older 0.5.6 OLSRs do not run well longer than 25 or 50 days
> > (depending on the version) because of an integer overflow in the timers.
> > 
> > The bug was fixed in 0.5.6-r6.
> 
> Was the bug fix so that olsr will not start a duplicate or so that it
> would not propagate it? I assume only to not start it because I see the
> propagation?
> 
> It is happening again. This time I see two duplicated hna messages.
> Currently the only way I can get rid of these is by blocking olsr packets
> from the originator address.
> 
> Would it be a mission to add code to filter duplicates? Is there a use
> for duplicate messages? Up to now I have only seen hna duplicate messages.
> Would there be a reason for it? I guess the place to start is hna_set.c?
OLSR contains already generic code to filter duplicates. See 
src/duplicate_set.c

Henning Rogge

-- 
1) You can't win.
2) You can't break even.
3) You can't leave the game.
— The Laws of Thermodynamics, summarized
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.olsr.org/pipermail/olsr-users/attachments/20100908/244b910e/attachment.sig>


More information about the Olsr-users mailing list