Re: Specifying interface for iperf
A couple of things...
cat /proc/net/dev_mcast from another window just before and just after
running iperf. That will tell you for sure which device the kernel is
associating the multicast group with.
You can also tcpdump -i eth1 to see the packets. (There was actually quite
a long delay before I saw IGMP packets -- say 10 seconds. This may be
normal, or it may be because I didn't do a "-n" and some nameserver issue
had to time out...)
Also, I just checked the syntax on the route command below and the "-dev"
option barfed. (The correct syntax is to not use it at all.) Make sure
the route is actually getting created correctly. (i.e. run the "route"
command).
I just repeated all of this on my machine, and it worked. The only other
thing I can think of is that maybe ip_forwarding has to be enabled to do
this. (I don't *think* so, but maybe...)
Jim
cnelson --at-- nycap.rr.com wrote:
>
> Thanks for the note.
>
> > Chris,
> > I posted a few months back with the same kind of problem -- doing
> > multicast testing using clients/servers with multiple network
> > interface cards...
> >
> > As an immediate fix for you, you should set up a static route on your
> > server machine so that the kernel knows to send the join requests
> > to the
> > proper interface. (Otherwise it just picks the first one on the list,
> > which in your case is probably the "other one.")
> >
> > Something like this should work:
> >
> > route add -net 224.0.0.0 netmask 255.225.0.0 -dev eth1
> >
> > then any multicast address 224.0.*.* would be "bound" to eth1.
> ...
>
> Nope, that didn't seem to do it. <shrug>
>
> Search the archives for a posting from me on this
> > topic to
> > get more details.
>
> A fine idea. If I could find the archives. I poked around http://dast.nlanr.net/Projects/Iperf/ without any success. Searching for "mailing list" and "archive" didn't help, either.
--
----------------------------------------------------------------------
James T. Kaba
Sarnoff Corporation There are 10 kinds of people in the world:
jkaba --at-- sarnoff.com those who understand binary, and those who don't.
609-734-2246