RE: [mpls-linux-general] label stacking with LDP in zebra
Status: Beta
Brought to you by:
jleu
|
From: Markus H. <mh...@as...> - 2003-03-07 11:44:11
|
Hi all,
i found a issue with MPLS Zebra-mplsd (all are newest versions from =
CVS):
I use a small testsetup with 4 LSR, routed by OSPF (all area 0) and LDP =
works, but /proc/net/mpls_out reports garbage:
0x00000003 2200/136388/0 5 PUSH(gen 10032) SET(eth2,10.3.6.3)=20
because eth2 is:
LSR-6:/etc/rc.d$ ifconfig eth2
eth2 Link encap:Ethernet HWaddr 00:00:0A:06:0A:06 =20
inet addr:10.6.10.6 Bcast:10.6.10.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:22792 errors:0 dropped:0 overruns:0 frame:0
TX packets:11316 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100=20
RX bytes:1577487 (1.5 Mb) TX bytes:947570 (925.3 Kb)
Interrupt:5=20
and eth1 is the correct one:
LSR-6:/etc/rc.d$ ifconfig eth1
eth1 Link encap:Ethernet HWaddr 00:00:0A:03:06:06 =20
inet addr:10.3.6.6 Bcast:10.3.6.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:28146 errors:0 dropped:0 overruns:0 frame:0
TX packets:10651 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100=20
RX bytes:1956041 (1.8 Mb) TX bytes:889563 (868.7 Kb)
Interrupt:5=20
the IP 10.3.6.3 is the correct IP of the next LSR
and all MPLS packets are correct routed to eth1.
the device for the proc output is initialized here (mpsl_utils.c:110):
static struct mpls_dst *mpls_make_dst(unsigned int ifi,struct sockaddr =
*nh,
struct mpls_out_info *moi) {
struct net_device *dev =3D NULL;
struct mpls_dst *md =3D NULL;
struct rtable *rt =3D NULL;
unsigned int nhip;
MPLS_DEBUG(("mpls_make_dst: enter\n"));
if (!(dev =3D dev_get_by_index(ifi))) {
goto mpls_make_dst_4;
}
maybe 'ifi' is wrong?
added entries with mplsadm2 are correct.
any ideas? more debug output needed ? (i can reproduce this every time)
btw, some answers for this question:
> Hi all,
>=20
> does anybody know a simple config for (zebra-)mplsd to generate
> a setup with a labelstack > 1 ?
>=20
?
thx in advance,
markus
|