Re: [mpls-linux-general] trouble setting up LSPs

[James R. L. <jl...@mi...>]

Hmmm.

Ok lets look at /proc/net/rt_cache  find the entries that match the
source and dest of the ftp session.

I'll need the info from both ends.

Also try running the transfer with mpls kernel debugging turned
on 'mplsadm2 -d'

Jim

On Wed, Sep 11, 2002 at 08:44:21PM -0400, Wayne W. Szeto wrote:
> 
> I checked the drop counter on the outlabels on all the machines and they
> are all zero.  The following are from /proc/net/mpls* on each machine:
> 
> 
> On A:
> =====
> [wwszeto@csgradpc03 net]$ more mpls_*
> ::::::::::::::
> mpls_in
> ::::::::::::::
> 0x40005000 7533/10804580/0 gen 20 0 7534 POP PEEK
> ::::::::::::::
> mpls_labelspace
> ::::::::::::::
> eth0    0       12
> ::::::::::::::
> mpls_out
> ::::::::::::::
> 0x00000006 4126/219279/0 1 PUSH(gen 16) SET(eth0,10.0.0.20)
> ::::::::::::::
> mpls_tunnel
> ::::::::::::::
> [wwszeto@csgradpc03 net]$ ip2 route
> 10.0.1.30 via 10.0.0.20 dev eth0 lsp 0x6
> 10.0.0.0/24 dev eth0  scope link
> 10.0.1.0/24 dev eth2  scope link
> 129.97.78.0/24 dev eth1  scope link
> 10.1.0.0/24 dev eth3  scope link
> 127.0.0.0/8 dev lo  scope link
> default dev eth1  scope link
> 
> On B:
> =====
> [root@node02 net]# more mpls*
> ::::::::::::::
> mpls_in
> ::::::::::::::
> 0x40004000 4126/235805/0 gen 16 0 1 POP FWD(0x00000002)
> 0x40005000 7533/10804580/0 gen 20 0 1 POP FWD(0x00000003)
> ::::::::::::::
> mpls_labelspace
> ::::::::::::::
> eth0    0       9
> eth1    0       8
> ::::::::::::::
> mpls_out
> ::::::::::::::
> 0x00000002 4126/219301/0 2 PUSH(gen 16) SET(eth1,10.0.1.30)
> 0x00000003 7533/10774448/0 2 PUSH(gen 20) SET(eth0,10.0.0.10)
> ::::::::::::::
> mpls_tunnel
> ::::::::::::::
> [root@node02 net]# ip2 route
> 10.0.0.0/24 dev eth0  scope link
> 10.0.1.0/24 dev eth1  scope link
> 10.1.0.0/24 dev eth2  scope link
> 127.0.0.0/8 dev lo  scope link
> default via 10.0.0.1 dev eth0
> 
> On C:
> =====
> [root@node03 net]# more mpls*
> ::::::::::::::
> mpls_in
> ::::::::::::::
> 0x40004000 4126/265835/0 gen 16 0 4127 POP PEEK
> ::::::::::::::
> mpls_labelspace
> ::::::::::::::
> eth1    0       9
> ::::::::::::::
> mpls_out
> ::::::::::::::
> 0x00000003 7575/10783790/0 1 PUSH(gen 20) SET(eth1,10.0.1.20)
> ::::::::::::::
> mpls_tunnel
> ::::::::::::::
> [root@node03 net]# ip2 route
> 10.0.0.10 via 10.0.1.20 dev eth1 lsp 0x3
> 10.0.0.0/24 dev eth0  scope link
> 10.0.1.0/24 dev eth1  scope link
> 10.1.0.0/24 dev eth2  scope link
> 127.0.0.0/8 dev lo  scope link
> default via 10.0.0.1 dev eth0
> 
> 
> On Wed, 11 Sep 2002, James R. Leu wrote:
> 
> > OK.  So the bug must still exist.  What are the symptoms you see?
> > Are there any kernel opps?  My guess is that mpls-linux is dropping
> > the frames because it thinks they are too big (>MTU).  So you should
> > see the drop counter increasing on an outlabel somewhere in the network.
> > (in /proc/net/mpls_out their is a trio of numbers for each entry in the
> > form 0/0/0, drops is the last number)
> >
> > Jim
> >

-- 
James R. Leu