Date: Mon, 11 Jan 1999 15:34:39 -0500 From: "Donald J . Maddox" <dmaddox@conterra.com> To: multimedia@FreeBSD.ORG Subject: How to do MBONE over a point-to-point connection? Message-ID: <19990111153439.A2156@dmaddox.conterra.com>
next in thread | raw e-mail | index | archive | help
I'm not 100% certain that this is the right forum for
this question, but I know that a lot of you guys on the
multimedia list have experience with MBONE stuff, so here
goes...
I have a simple 128K ISDN dial-up PPP conection to my ISP.
My ISP was kind enough to set up a multicast tunnel for me on
their multicast gateway machine at my request. They have an
/etc/mrouted.conf containing the line:
tunnel 209.12.169.30 209.12.169.48 metric 1 threshold 1 rate_limit 120 passive rexmit_prunes off noflood
on the multicast gateway machine. I initially thought that
I could simply create an /etc/mrouted.conf on my machine
containing only the line:
tunnel tun0 209.12.169.30 metric 1 advert_metric 20 threshold 1 rate_limit 120
and I would have access to the MBONE. Unfortunately, it doesn't
appear to be that simple.
Mrouted does not like point-to-point connections, and is
not designed to work with such. Normally, my connection to
my isp is 209.12.169.48->209.12.169.2 netmask 0xffffffff.
Netmasks don't mean much on PTP interfaces, but they mean a
lot to mrouted. Mrouted doesn't believe any netmask more
restrictive than 0xffffff00 is a valid subnet, and will refuse
to operate on an interface with such a netmask. I thought I
could get around this by simply setting the netmask to
0xffffff00 (since it has no effect on a PTP connection), but,
unfortunately, this cannot work, because the IP address of
the MBONE gateway is 209.12.169.30, and this netmask puts
us on the same subnet. Mrouted then refuses to create an
"unnecessary tunnel" between two hosts on the same subnet :-(
I worked around this by simply hacking the mrouted
source to ignore the fact that both ends of the tunnel were
on the same subnet.
This appeared to work, more or less... Unfortunately,
mrouted automatically configures itself to use ALL multicast-
capable interfaces, so it ended up sending all traffic over
the 'tun0' interface directly, instead of using the tunnel.
So... I tried changing my mrouted.conf to this:
phyint tun0 disable
tunnel tun0 209.12.169.30 metric 1 advert_metric 20 threshold 1 rate_limit 120
Mrouted then refused to start, because it MUST have at least
two enabled VIFs. So... I changed my mrouted.conf again:
phyint tun0 disable
tunnel tun0 209.12.169.30 metric 1 advert_metric 20 threshold 1 rate_limit 120
tunnel tun0 1.2.3.4 metric 1 advert_metric 30 threshold 1 rate_limit 80
and set up a static route 1.2.3.4->127.0.0.1, so mrouted _thinks_
it has two tunnels.
Now, mrouted starts just fine and I get _lots_ of multicast
traffic over the tunnel, mostly routing info. So much, in fact,
that it more or less swamped my rather puny bandwidth. I have
found, though, that setting my 'advert_metric' to 20 or so at
least keeps me from getting flooded.
Unfortuneately, it _still_ doesn't seem to work as expected.
Using tcpdump, I can see lots of multicast traffic between my
machine and the multicast gateway machine, but if I start sdr,
I see my end send packets to SAP.MCAST.NET or ALL-ROUTERS.MCAST.NET,
followed immediately by ICMP ttl-exceeded messages from the dial-up
router, as if these packets are somehow not getting routed through
the tunnel.
At this point, I am just about out of ideas... It certainly
_seems_ that what I am trying to do should be possible. Anybody
got any tips for me?
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-multimedia" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19990111153439.A2156>