On Nov 24, 2015, at 6:52 PM, Shen Ting Ang <angshent...@gmail.com> wrote:

> Hi all,
> 
> Following up from my previous question about POX and fat-tree topologies, 
> I've tried using forwarding.l2multi and openflow.discovery. However, I now 
> run into a slightly different problem: there's only connectivity between 
> several hosts when I run pingall on my Mininet CLI.

I think it should be possible to make this work, but I don't think I've tried 
it recently if ever.  It may take some tweaking of parameters.  I've posted 
some suggested parameters to start with for discovery/spanning_tree on the 
mailing list and possibly in the POX manual.  What are the parameters you're 
using, if any?

> I've also looked up RipL-POX and have tried running it, but it seems that the 
> code was meant to run on an older version of POX and I'm getting all kinds of 
> errors just following the default instructions listed in the github. Or 
> should I find an older version of POX that is RipL compatible...?

That's one possibility.  I also did some work on compatibility a while back.  
If I recall correctly, I forked RipL-POX and I think there's an issue or pull 
request against the original RipL or RipL-POX which describes some of my 
experiences getting things to work with more recent Mininet/POX.  There may 
also be some notes in the POX wiki (I can't remember).

So yeah.  You could go oldschool, but that probably means digging up both old 
Mininet and old POX versions.  Or you could try to start from the work I did a 
year or two ago on getting things running on recent versions again.  I'd do the 
latter (obviously), but I think at least one person has had some semi-recent 
success with the former approach too.

-- Murphy

Reply via email to