mrouted and IGMP on Linux

I've got two Linux machines (Debian Wally) both running mrouted.
WinPC1 is sending multicast traffic and WinPC2 is receiving the
multicast traffic.

WinPC1 -> DebianA -> DebianB -> WinPC2

My problem seems to occur on DebianB. If I "tcpdump -i eth0 ip
multicast" I can see the packets coming in, so DebianA properly routed
to DebianB using mrouted:

10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
...

And if I "tcpdump -i eth1 ip multicast" I can see the join coming in
from WinPC2 for the appropriate IGMP group:

192.168.0.4 > 239.255.12.42: igmp v2 report 239.255.12.42 [ttl 1]
and later on...
192.168.0.4 > 224.0.0.2: igmp leave 239.255.12.42 [ttl 1]

I have verified that the traffic ttl is high enough (12 actually).
And a check of 'mrinfo' looks good:

127.0.0.1 (debianb) [DVMRPv3 compliant]:
182.158.0.7 -> 182.158.0.9 (182.158.0.9) [1/1/querier]
192.168.0.6 -> 0.0.0.0 (local) [1/1/querier/leaf]

Strangely this works when I remove one of the two Debian machines from
the chain. So my guess is that 'mrouted' is getting the packet
downstream from debiana to debianb, but then interferring on debianb
where that actual 'join' is coming in from WinPC2.

Thanks,
M

In the past I have tried (without success) to get an identical configuration
to work using mrouted. I decided to use smcroute instead of mrouted for my
application, but I still wonder why mrouted won't work.
I'd be interested in taking a look at the routing tables on Debian A and
Debian B.

"Matt" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) om...
> I've got two Linux machines (Debian Wally) both running mrouted.
> WinPC1 is sending multicast traffic and WinPC2 is receiving the
> multicast traffic.
>
> WinPC1 -> DebianA -> DebianB -> WinPC2
>
> My problem seems to occur on DebianB. If I "tcpdump -i eth0 ip
> multicast" I can see the packets coming in, so DebianA properly routed
> to DebianB using mrouted:
>
> 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> ...
>
> And if I "tcpdump -i eth1 ip multicast" I can see the join coming in
> from WinPC2 for the appropriate IGMP group:
>
> 192.168.0.4 > 239.255.12.42: igmp v2 report 239.255.12.42 [ttl 1]
> and later on...
> 192.168.0.4 > 224.0.0.2: igmp leave 239.255.12.42 [ttl 1]
>
> I have verified that the traffic ttl is high enough (12 actually).
> And a check of 'mrinfo' looks good:
>
> 127.0.0.1 (debianb) [DVMRPv3 compliant]:
> 182.158.0.7 -> 182.158.0.9 (182.158.0.9) [1/1/querier]
> 192.168.0.6 -> 0.0.0.0 (local) [1/1/querier/leaf]
>
> Strangely this works when I remove one of the two Debian machines from
> the chain. So my guess is that 'mrouted' is getting the packet
> downstream from debiana to debianb, but then interferring on debianb
> where that actual 'join' is coming in from WinPC2.
>
> Thanks,
> M

Here are the dumps from mrouted from debiana and debianb...

Thanks,
Matt

--------------------------

127.0.0.1 (debiana) [DVMRPv3 compliant]:
10.3.30.249 -> 0.0.0.0 (local) [1/1/querier/leaf]
182.158.0.9 -> 182.158.0.8 (182.158.0.8) [1/1]

mrouted version 3.9-beta3 up 6:13:04 Thu Aug 21 06:13:04 2003

vifs_with_neighbors = 1
[This host is a leaf]


Virtual Interface Table
Vif Name Local-Address M Thr Rate
Flags
0 wancom0 10.3.30.249 subnet: 10/9 1 1 0
querier leaf
group host (time left): 224.0.1.60 10.3.30.19 (
0:04:09)
235.80.68.83 10.3.30.109 (
0:04:08)
IGMP querier: 10.3.30.249 (this system)
Nbr bitmaps: 0x0000000000000000

1 wancom1 182.158.0.9 subnet: 182/9 1 1 0
peers: 182.158.0.8 (3.255) [0] have-genid
up 1:22:58
group host (time left): 224.0.0.4 182.158.0.8 (
0:03:28)
224.0.0.2 182.158.0.8 (
0:03:31)
IGMP querier: 182.158.0.8 up 6:11:51 last
heard 0:00:50 ago
Nbr bitmaps: 0x0000000000000001


Multicast Routing Table (3 entries)
Origin-Subnet From-Gateway Metric Tmr Fl In-Vif Out-Vifs
192.168.0/25 182.158.0.8 2 50 .. 1 0*
182/9 1 190 .. 1 0*
10/9 1 190 .. 0 1[0]

mrouted version 3.9-beta3 up 6:13:04 Thu Aug 21 06:13:04 2003

Multicast Routing Cache Table (1 entries)
Origin Mcast-group CTmr Age Ptmr Rx IVif
Forwvifs
<(prunesrc:vif[idx]/tmr) prunebitmap
>Source Lifetime SavPkt Pkts Bytes RPFf
10/9 239.255.12.42 0:01:37 0:02:06 - - 0
1
>10.3.30.104 0:02:06 0

--------------------------

127.0.0.1 (debianb) [DVMRPv3 compliant]:
182.158.0.8 -> 182.158.0.9 (182.158.0.9) [1/1/querier]
192.168.0.5 -> 0.0.0.0 (local) [1/1/querier/leaf]


mrouted version 3.9-beta3 up 1:21:01 Thu Aug 21 10:35:42 2003

vifs_with_neighbors = 1
[This host is a leaf]


Virtual Interface Table
Vif Name Local-Address M Thr Rate
Flags
0 eth0 182.158.0.8 subnet: 182.158.0/24 1 1 0
querier
peers: 182.158.0.9 (3.255) [0] have-genid
up 1:21:01
group host (time left): 224.0.0.4 182.158.0.8 (
0:04:02)
224.0.0.2 182.158.0.8 (
0:04:03)
IGMP querier: 182.158.0.8 (this system)
Nbr bitmaps: 0x0000000000000001
pkts/bytes in : 0/0
pkts/bytes out: 0/0

1 eth1 192.168.0.5 subnet: 192.168.0/24 1 1 0
querier leaf
group host (time left): 239.255.12.42 192.168.0.4 (
0:04:15)
235.80.68.83 192.168.0.4 (
0:02:24)
224.0.0.4 192.168.0.5 (
0:02:31)
224.0.0.2 192.168.0.5 (
0:02:30)
IGMP querier: 192.168.0.5 (this system)
Nbr bitmaps: 0x0000000000000000
pkts/bytes in : 0/0
pkts/bytes out: 0/0


Multicast Routing Table (3 entries)
Origin-Subnet From-Gateway Metric Tmr Fl In-Vif Out-Vifs
192.168.0/24 1 70 .. 1 0[0]
182.158.0/24 1 70 .. 0 1*
10/8 182.158.0.9 2 55 .. 0 1*

mrouted version 3.9-beta3 up 1:21:01 Thu Aug 21 10:35:42 2003

Multicast Routing Cache Table (0 entries)
Origin Mcast-group CTmr Age Ptmr Rx IVif
Forwvifs
<(prunesrc:vif[idx]/tmr) prunebitmap
>Source Lifetime SavPkt Pkts Bytes RPFf

--------------------------


"ERG" <(E-Mail Removed)> wrote in message news:<wNg1b.108410$(E-Mail Removed)>...
> In the past I have tried (without success) to get an identical configuration
> to work using mrouted. I decided to use smcroute instead of mrouted for my
> application, but I still wonder why mrouted won't work.
> I'd be interested in taking a look at the routing tables on Debian A and
> Debian B.
>
> "Matt" <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed) om...
> > I've got two Linux machines (Debian Wally) both running mrouted.
> > WinPC1 is sending multicast traffic and WinPC2 is receiving the
> > multicast traffic.
> >
> > WinPC1 -> DebianA -> DebianB -> WinPC2
> >
> > My problem seems to occur on DebianB. If I "tcpdump -i eth0 ip
> > multicast" I can see the packets coming in, so DebianA properly routed
> > to DebianB using mrouted:
> >
> > 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> > 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> > 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> > ...
> >
> > And if I "tcpdump -i eth1 ip multicast" I can see the join coming in
> > from WinPC2 for the appropriate IGMP group:
> >
> > 192.168.0.4 > 239.255.12.42: igmp v2 report 239.255.12.42 [ttl 1]
> > and later on...
> > 192.168.0.4 > 224.0.0.2: igmp leave 239.255.12.42 [ttl 1]
> >
> > I have verified that the traffic ttl is high enough (12 actually).
> > And a check of 'mrinfo' looks good:
> >
> > 127.0.0.1 (debianb) [DVMRPv3 compliant]:
> > 182.158.0.7 -> 182.158.0.9 (182.158.0.9) [1/1/querier]
> > 192.168.0.6 -> 0.0.0.0 (local) [1/1/querier/leaf]
> >
> > Strangely this works when I remove one of the two Debian machines from
> > the chain. So my guess is that 'mrouted' is getting the packet
> > downstream from debiana to debianb, but then interferring on debianb
> > where that actual 'join' is coming in from WinPC2.
> >
> > Thanks,
> > M

What do the unicast routing tables (route command output) on DebA and DebB
look like? For the config to work, I would expect each table to look
similar. In other words, Debian B's table should look like Debian A's table
except Debian A eth0 IP address would be replaced with the DebB eth0 IP and
the place where Debian A eth1 IP appears would be replaced with the DebianB
eth1 ip.

"Matt" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) om...
> Here are the dumps from mrouted from debiana and debianb...
>
> Thanks,
> Matt
>
> --------------------------
>
> 127.0.0.1 (debiana) [DVMRPv3 compliant]:
> 10.3.30.249 -> 0.0.0.0 (local) [1/1/querier/leaf]
> 182.158.0.9 -> 182.158.0.8 (182.158.0.8) [1/1]
>
> mrouted version 3.9-beta3 up 6:13:04 Thu Aug 21 06:13:04 2003
>
> vifs_with_neighbors = 1
> [This host is a leaf]
>
>
> Virtual Interface Table
> Vif Name Local-Address M Thr Rate
> Flags
> 0 wancom0 10.3.30.249 subnet: 10/9 1 1 0
> querier leaf
> group host (time left): 224.0.1.60 10.3.30.19 (
> 0:04:09)
> 235.80.68.83 10.3.30.109 (
> 0:04:08)
> IGMP querier: 10.3.30.249 (this system)
> Nbr bitmaps: 0x0000000000000000
>
> 1 wancom1 182.158.0.9 subnet: 182/9 1 1 0
> peers: 182.158.0.8 (3.255) [0] have-genid
> up 1:22:58
> group host (time left): 224.0.0.4 182.158.0.8 (
> 0:03:28)
> 224.0.0.2 182.158.0.8 (
> 0:03:31)
> IGMP querier: 182.158.0.8 up 6:11:51 last
> heard 0:00:50 ago
> Nbr bitmaps: 0x0000000000000001
>
>
> Multicast Routing Table (3 entries)
> Origin-Subnet From-Gateway Metric Tmr Fl In-Vif Out-Vifs
> 192.168.0/25 182.158.0.8 2 50 .. 1 0*
> 182/9 1 190 .. 1 0*
> 10/9 1 190 .. 0 1[0]
>
> mrouted version 3.9-beta3 up 6:13:04 Thu Aug 21 06:13:04 2003
>
> Multicast Routing Cache Table (1 entries)
> Origin Mcast-group CTmr Age Ptmr Rx IVif
> Forwvifs
> <(prunesrc:vif[idx]/tmr) prunebitmap
> >Source Lifetime SavPkt Pkts Bytes RPFf
> 10/9 239.255.12.42 0:01:37 0:02:06 - - 0
> 1
> >10.3.30.104 0:02:06 0
>
> --------------------------
>
> 127.0.0.1 (debianb) [DVMRPv3 compliant]:
> 182.158.0.8 -> 182.158.0.9 (182.158.0.9) [1/1/querier]
> 192.168.0.5 -> 0.0.0.0 (local) [1/1/querier/leaf]
>
>
> mrouted version 3.9-beta3 up 1:21:01 Thu Aug 21 10:35:42 2003
>
> vifs_with_neighbors = 1
> [This host is a leaf]
>
>
> Virtual Interface Table
> Vif Name Local-Address M Thr Rate
> Flags
> 0 eth0 182.158.0.8 subnet: 182.158.0/24 1 1 0
> querier
> peers: 182.158.0.9 (3.255) [0] have-genid
> up 1:21:01
> group host (time left): 224.0.0.4 182.158.0.8 (
> 0:04:02)
> 224.0.0.2 182.158.0.8 (
> 0:04:03)
> IGMP querier: 182.158.0.8 (this system)
> Nbr bitmaps: 0x0000000000000001
> pkts/bytes in : 0/0
> pkts/bytes out: 0/0
>
> 1 eth1 192.168.0.5 subnet: 192.168.0/24 1 1 0
> querier leaf
> group host (time left): 239.255.12.42 192.168.0.4 (
> 0:04:15)
> 235.80.68.83 192.168.0.4 (
> 0:02:24)
> 224.0.0.4 192.168.0.5 (
> 0:02:31)
> 224.0.0.2 192.168.0.5 (
> 0:02:30)
> IGMP querier: 192.168.0.5 (this system)
> Nbr bitmaps: 0x0000000000000000
> pkts/bytes in : 0/0
> pkts/bytes out: 0/0
>
>
> Multicast Routing Table (3 entries)
> Origin-Subnet From-Gateway Metric Tmr Fl In-Vif Out-Vifs
> 192.168.0/24 1 70 .. 1 0[0]
> 182.158.0/24 1 70 .. 0 1*
> 10/8 182.158.0.9 2 55 .. 0 1*
>
> mrouted version 3.9-beta3 up 1:21:01 Thu Aug 21 10:35:42 2003
>
> Multicast Routing Cache Table (0 entries)
> Origin Mcast-group CTmr Age Ptmr Rx IVif
> Forwvifs
> <(prunesrc:vif[idx]/tmr) prunebitmap
> >Source Lifetime SavPkt Pkts Bytes RPFf
>
> --------------------------
>
>
> "ERG" <(E-Mail Removed)> wrote in message
news:<wNg1b.108410$(E-Mail Removed)>...
> > In the past I have tried (without success) to get an identical
configuration
> > to work using mrouted. I decided to use smcroute instead of mrouted for
my
> > application, but I still wonder why mrouted won't work.
> > I'd be interested in taking a look at the routing tables on Debian A and
> > Debian B.
> >
> > "Matt" <(E-Mail Removed)> wrote in message
> > news:(E-Mail Removed) om...
> > > I've got two Linux machines (Debian Wally) both running mrouted.
> > > WinPC1 is sending multicast traffic and WinPC2 is receiving the
> > > multicast traffic.
> > >
> > > WinPC1 -> DebianA -> DebianB -> WinPC2
> > >
> > > My problem seems to occur on DebianB. If I "tcpdump -i eth0 ip
> > > multicast" I can see the packets coming in, so DebianA properly routed
> > > to DebianB using mrouted:
> > >
> > > 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> > > 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> > > 10.3.30.104.4941 > 239.255.12.42.1234: udp 1316
> > > ...
> > >
> > > And if I "tcpdump -i eth1 ip multicast" I can see the join coming in
> > > from WinPC2 for the appropriate IGMP group:
> > >
> > > 192.168.0.4 > 239.255.12.42: igmp v2 report 239.255.12.42 [ttl 1]
> > > and later on...
> > > 192.168.0.4 > 224.0.0.2: igmp leave 239.255.12.42 [ttl 1]
> > >
> > > I have verified that the traffic ttl is high enough (12 actually).
> > > And a check of 'mrinfo' looks good:
> > >
> > > 127.0.0.1 (debianb) [DVMRPv3 compliant]:
> > > 182.158.0.7 -> 182.158.0.9 (182.158.0.9) [1/1/querier]
> > > 192.168.0.6 -> 0.0.0.0 (local) [1/1/querier/leaf]
> > >
> > > Strangely this works when I remove one of the two Debian machines from
> > > the chain. So my guess is that 'mrouted' is getting the packet
> > > downstream from debiana to debianb, but then interferring on debianb
> > > where that actual 'join' is coming in from WinPC2.
> > >
> > > Thanks,
> > > M

I've gotten feedback from someone who got this to work that it's not a
configuration issue, but a Linux issue that required source changes to
fix. I've since replaced debianb with a BSD 4.8 box and it works just
fine.

The issue is only in the "destination" LAN i.e. same box running
mrouted and receiving the relevant group joins. The Linux mrouted
appears to work anywhere else in the chain ("source" LAN or
inbetween.)

Matt