mpls-linux-general — Mailing list for general discussion about MPLS for Linux

Re: [mpls-linux-general] Iptables Extension to MPLS-LINUX

[Vincent J. <ja...@en...>]

1/ bug in ip_output.c:ip_output()

#ifdef CONFIG_MPLS_INGRESS_POLICING
        if (skb->mpls_index)
                return mpls_output(skb);
#endif

should be after 

#ifdef CONFIG_IP_ROUTE_NAT
        struct rtable *rt = (struct rtable*)skb->dst;
#endif

2/ bug in libipt_MPLS.c
static void
save(const struct ipt_ip *ip, const struct ipt_entry_target *target)
{
        const struct ipt_mpls_target_info *mplsinfo =
                (const struct ipt_mpls_target_info *)target->data;
 
        printf("--set-mpls 0x%lx ", mplsinfo->index);
} 
It should be --set-index
Moreover, there is the same error in your README, set-index should be used 
instead of set-mpls. Or you should update libipt_MPLS.c and replace the 
parsing of all the set-index with set-mpls. Otherwise it is not consistent.

3/ it does not work with my testbed. I am still workin on it ;-(

Vincent

PS:
int ip_output(struct sk_buff *skb)
{
 
#ifdef CONFIG_IP_ROUTE_NAT
        struct rtable *rt = (struct rtable*)skb->dst;
#endif
 
/* OD & AD : call mpls_output routine if mpls_index has been set in PREROUTING
 * or LOCAL_OUT hook both by netfilter or tc ingress.
 */
#ifdef CONFIG_MPLS_INGRESS_POLICING
        if (skb->mpls_index)
                return mpls_output(skb);
#endif
 
        IP_INC_STATS(IpOutRequests);

PPS: Please, could you post a more recent diff ?
PPPS: I am sorry if this mail is delayed, we have some routing loops in our 
local network ;-(

Le Lundi  2 Juillet 2001 10:39, Olivier Dugeon a écrit :
> Hi Vincent,
>
> Vincent Jardin wrote:
> > I just see a tiny bug in your README:
> >
> > You wrote:
> > ...
> > LER A:
> > ------
> >
> > LER A acts as usual with binding a fec to the outgoing label and
> > doesn't use iptables.
> >
> > # Add a route to LER B
> > #
> > route add -host 10.1.0.1 gw 10.1.0.2
> > ...
> > It should be
> > route add -host 10.2.1.2 gw 10.1.0.2
> > Otherwise you cannot add the LER A's FEC:
> > mplsadm -v -A -B -O gen:16:eth0:ipv4:10.1.0.2 -f 10.2.1.2/32
>
> Exact. we have modify the README. We found another tiny bug about the
> route.
>
> We wrote :
>
> The hosts are MPLS unaware. They just need to add an
> explicit route to the other Host like this :
>
> # on Host A
> route add -net 10.2.1.0/24 gw 10.1.1.2
>                                      ^
> # on Host B
> route add -net 10.1.1.0/24 gw 10.2.1.2
>                                      ^
> It should be
>
> # on Host A
> route add -net 10.2.1.0/24 gw 10.1.1.1
>                                      ^
> # on Host B
> route add -net 10.1.1.0/24 gw 10.2.1.1
>                                      ^
>
> Olivier & Alexandre

Re: [mpls-linux-general] Iptables Extension to MPLS-LINUX

[Vincent J. <ja...@en...>]

Hi,

I am trying to apply your patch, but it has not been created against the last 
release of the MPLS-linux patch.

Then, there are the following issues:
patching file include/linux/netfilter_ipv4/ipt_MPLS.h
patching file include/linux/netfilter_ipv4/ipt_mpls.h
patching file net/ipv4/ip_output.c
patching file net/ipv4/netfilter/Config.in
Hunk #1 succeeded at 18 (offset -1 lines).
Hunk #2 succeeded at 61 with fuzz 1 (offset -11 lines).
patching file net/ipv4/netfilter/Makefile
Hunk #1 succeeded at 48 (offset -5 lines).
patching file net/ipv4/netfilter/ipt_MPLS.c
patching file net/ipv4/netfilter/ipt_mpls.c
patching file net/ipv4/route.c
patch: **** malformed patch at line 257: diff -uNr 
linux-2.4.5/net/mpls/mpls_output.c 
linux-2.4.5-mpls_ingress/net/mpls/mpls_output.c

Please, could you create a more recent patch ?

Thanks,
  Vincent

Re: [mpls-linux-general] Iptables Extension to MPLS-LINUX

[Olivier D. <Oli...@rd...>]

Hi Vincent,

Vincent Jardin wrote:
> 
> I just see a tiny bug in your README:
> 
> You wrote:
> ...
> LER A:
> ------
> 
> LER A acts as usual with binding a fec to the outgoing label and
> doesn't use iptables.
> 
> # Add a route to LER B
> #
> route add -host 10.1.0.1 gw 10.1.0.2
> ...
> It should be
> route add -host 10.2.1.2 gw 10.1.0.2
> Otherwise you cannot add the LER A's FEC:
> mplsadm -v -A -B -O gen:16:eth0:ipv4:10.1.0.2 -f 10.2.1.2/32
> 

Exact. we have modify the README. We found another tiny bug about the
route.

We wrote :

The hosts are MPLS unaware. They just need to add an
explicit route to the other Host like this :

# on Host A
route add -net 10.2.1.0/24 gw 10.1.1.2
                                     ^
# on Host B
route add -net 10.1.1.0/24 gw 10.2.1.2
                                     ^
It should be

# on Host A
route add -net 10.2.1.0/24 gw 10.1.1.1
                                     ^
# on Host B
route add -net 10.1.1.0/24 gw 10.2.1.1
                                     ^

Olivier & Alexandre
-- 
 FTR&D/DAC/CPN
 Technopole Anticipa     | mailto:Oli...@fr...
 2, Avenue Pierre Marzin | Phone:  +(33) 2 96 05 28 80
 F-22307 LANNION         | Fax:    +(33) 2 96 05 18 52

Re: [mpls-linux-general] Iptables Extension to MPLS-LINUX

[Vincent J. <ja...@en...>]

I just see a tiny bug in your README:

You wrote:
...
LER A:
------
 
LER A acts as usual with binding a fec to the outgoing label and
doesn't use iptables.
 
# Add a route to LER B
#
route add -host 10.1.0.1 gw 10.1.0.2
...
It should be
route add -host 10.2.1.2 gw 10.1.0.2
Otherwise you cannot add the LER A's FEC:
mplsadm -v -A -B -O gen:16:eth0:ipv4:10.1.0.2 -f 10.2.1.2/32

Vincent

Le Vendredi 29 Juin 2001 05:23, Olivier Dugeon a écrit :

> > Hi Jim,
>
> Here it is a patch to use iptables/netfilter on top of mpls-linux. It
> let you mark and label packet using all the power of iptables filter.
>
> Let us know if you have problems.
>
> Have fun with it,
>
> Olivier and Alexandre

----------------------------------------
Content-Type: application/octet-stream; charset="us-ascii"; 
name="mpls-iptables-0.2.tgz"
Content-Transfer-Encoding: base64
Content-Description: 
----------------------------------------

Re: [mpls-linux-general] another problem in ldp_linux

[Vincent J. <ja...@en...>]

>
> LDP will only install out labels and bind them to a FEC if the FEC has an
> EXACT match in the routin table.  While in ldp_linux do a 'show route'.
> Every LDP speaker should have the same entries.
>
> So to get LSPs to the router IDs of each LDP speaker, each LDP speaker
> needs to have a /32 route to the router ID of the other LDP speakers.


I tried it, but I still had the same problem, the FEC remained empty.

Please, could you post us a simple example for which it is working ?

Thanks,
  Vincent

Re: [mpls-linux-general] another problem in ldp_linux

[Vincent J. <ja...@en...>]

Thanks for your answer, but I still have the same problem. See the logs below.

> LDP will only install out labels and bind them to a FEC if the FEC has an
> EXACT match in the routin table.  While in ldp_linux do a 'show route'.
> Every LDP speaker should have the same entries.
>
> So to get LSPs to the router IDs of each LDP speaker, each LDP speaker
> needs to have a /32 route to the router ID of the other LDP speakers.


Hmm, I tried it with a /32 static route to the router but I could net get the 

(on each host, I run first zebra -d with the enclosed configuration that has 
a static /32 route to router ID, then mpldadm -L eth0:0, then ldp_linux)

FECs:
**** host 10.0.0.6: ****
bash-2.04# ip route show
10.0.0.5 via 192.168.12.1 dev eth0  proto zebra 
192.168.12.0/24 dev eth0  proto kernel  scope link  src 192.168.12.1 
127.0.0.0/8 dev lo  scope link 
bash-2.04# cat /proc/net/mpls_in
40004000 gen 16 0 POP DLV 
40004400 gen 17 0 POP DLV 
40004800 gen 18 0 POP DLV 
40004c00 gen 19 0 POP DLV 
bash-2.04# cat /proc/net/mpls_out
bash-2.04# cat /proc/net/mpls_fec
bash-2.04# cat /proc/net/mpls_labelspace
lo      0
teql0   0
eth0    0
bash-2.04# cat /etc/zebra/zebra.conf
! zebra
! $Release$, $Date$, $Author$
!
password zebra
!
interface eth0
  ip address 192.168.12.1/24
!
interface lo
  ip address 10.0.0.6/32
!
! static route to the router ID 10.0.0.5 through eth0
ip route 10.0.0.5/32 192.168.12.1
!
log file zebra.log
bash-2.04#

-----
prompt>add global 10.0.0.6
Adding interface lo
Adding interface eth0
Adding route 0a000005/32 via 6
Adding route c0a80c00/24 via 6
Adding route 7f000000/8 via 1
Adding route 0a000006/32 via 1
Adding global object with LSRID 0a000006
prompt>add interface eth0
Adding interface eth0
prompt>PRT: index:   1 0a000005/32 2
here
here
here
here
PRT: index:   6 0a000006/32 2
PRT: index:   7 7f000000/8 2
PRT: index:   8 c0a80c00/24 2

prompt>show route
0a000005/32 via c0a80c01:6
0a000006/32 via 00000000:1
7f000000/8 via 00000000:1
c0a80c00/24 via 00000000:6
prompt>


**** host 10.0.0.5 ****
bash-2.04# ip route show
10.0.0.6 via 192.168.12.2 dev eth0  proto zebra 
192.168.12.0/24 dev eth0  proto kernel  scope link  src 192.168.12.2 
127.0.0.0/8 dev lo  scope link 
bash-2.04# cat /proc/net/mpls_in
40004000 gen 16 0 POP DLV 
40004400 gen 17 0 POP DLV 
40004800 gen 18 0 POP DLV 
40004c00 gen 19 0 POP DLV 
bash-2.04# cat /proc/net/mpls_out
bash-2.04# cat /proc/net/mpls_fec
bash-2.04# cat /proc/net/mpls_labelspace 
lo      0
teql0   0
eth0    0
bash-2.04# cat /etc/zebra/zebra.conf
! zebra
! $Release$, $Date$, $Author$
!
password zebra
!
interface eth0
  ip address 192.168.12.2/24
!
interface lo
  ip address 10.0.0.5/32
!
! static route to the router ID 10.0.0.6 through eth0
ip route 10.0.0.6/32 192.168.12.2
!
log file zebra.log
bash-2.04# 

---
prompt>add global 10.0.0.5
Adding interface lo
Adding interface eth0
Adding route 0a000006/32 via 6
Adding route c0a80c00/24 via 6
Adding route 7f000000/8 via 1
Adding route 0a000005/32 via 1
Adding global object with LSRID 0a000005
prompt>add interface eth0
Adding interface eth0
prompt>here
here
here
here
PRT: index:   5 0a000005/32 2
PRT: index:   6 0a000006/32 2
PRT: index:   7 7f000000/8 2
PRT: index:   8 c0a80c00/24 2

prompt>




--
Vincent

[mpls-linux-general] Iptables Extension to MPLS-LINUX

[Olivier D. <Oli...@rd...>]

Hi Jim,

Here it is a patch to use iptables/netfilter on top of mpls-linux. It
let you mark and label packet using all the power of iptables filter.

Let us know if you have problems.

Have fun with it,

Olivier and Alexandre
-- 
 FTR&D/DAC/CPN
 Technopole Anticipa     | mailto:Oli...@fr...
 2, Avenue Pierre Marzin | Phone:  +(33) 2 96 05 28 80
 F-22307 LANNION         | Fax:    +(33) 2 96 05 18 52

Re: [mpls-linux-general] another problem in ldp_linux

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

On Thu, Jun 28, 2001 at 05:16:23PM -0400, Vincent Jardin wrote:
> I could get quite the same problem, but there is a slight difference. I=
 need=20
> to configure 10.0.0.6 first.
>=20
> When it does not work, I get the error:
>   PRT: ldp_state_recv_init: cannot find adj
> on the router 10.0.0.6 when 10.0.0.5 is configured first.
>=20
> Otherwise, the mpls_in is filled properly when I configure 10.0.0.6 fir=
st.
>=20
> In the case it works (10.0.0.6 is the first router that I configure):
> 10.0.0.5 :
> $ cat /proc/net/mpls_in =20
> 40004000 gen 16 0 POP DLV=20
> 40004400 gen 17 0 POP DLV=20
> 40004800 gen 18 0 POP DLV
> but mpls_out, and mpls_fec remain empty.
> Why ?

LDP will only install out labels and bind them to a FEC if the FEC has an
EXACT match in the routin table.  While in ldp_linux do a 'show route'.
Every LDP speaker should have the same entries.

So to get LSPs to the router IDs of each LDP speaker, each LDP speaker ne=
eds
to have a /32 route to the router ID of the other LDP speakers.

Jim

>=20
> 10.0.0.6:
> $ cat /proc/net/mpls_in =20
> 40004000 gen 16 0 POP DLV=20
> 40004400 gen 17 0 POP DLV=20
> 40004800 gen 18 0 POP DLV=20
> but mpls_out, and mpls_fec remain empty.
> Why ?
>=20
> Vincent
>=20
> Le Mercredi 27 Juin 2001 05:19, Venisa Cabrilla a =E9crit :
> > Jim,
> >
> > Sorry, another problem ... :-)
> >
> > I know that my problem here will seem somewhat ridiculous but I have =
spent
> > two and half hours just to find my mistakes ...
> >
> > Please refer to the same configuration I gave you before, but now wit=
h
> > mplsadm of course :
> >
> >
> >     lo 10.0.0.6                lo 10.0.0.5
> >      ---------                  ---------
> >
> >     | Router2 |----------------| Router1 |
> >
> >      ---------                  ---------
> >
> >          |eth2  192.168.12.0/24     |eth1
> >
> >     ---------------------------------------
> >
> > In router1:
> > bashH ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > bashH route add 10.0.0.6 gw 192.168.12.2
> > bashH mplsadm -L eth1:0
> >
> > >From inside ldp_linux:
> >
> > prompt> add global 10.0.0.5
> > prompt> add interface eth1
> >
> > In router2:
> > bashH ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > bashH route add 10.0.0.5 gw 192.168.12.1
> > bashH mplsadm -L eth2:0
> >
> > >From inside ldp_linux:
> >
> > prompt> add global 10.0.0.6
> > prompt> add interface eth2
> >
> > I do not know if the order of which we run and configure ldp_linux wi=
ll
> > certainly determine the result, but in my case, yes !
> >
> > There are two possible cases, we run and configure ldp_linux first in
> > router1 and then in router2, the other, we run and configure ldp_linu=
x in
> > router2 and then router1. What I mean by running and configuring ldp_=
linux
> > here is that we type add global ... and add interface ...
> >
> > First, we run and configure ldp_linux first in router1(10.0.0.5) and =
then
> > in router2 (10.0.0.6), ldp_linux works well, I can see that mpls_* ar=
e not
> > empty and I can ping the machines. I can see the label mapping proces=
s both
> > within ldp_linux or by using ethereal. Everything seems right on its =
path.
> >
> > Second, we run and configure ldp_linux first in router2(10.0.0.6) and=
 then
> > in router1(10.0.0.1), well here comes the problem, mpls_* rest empty =
and I
> > can only see hello messages using ethereal, nothing else.
> > There is no label mapping process and ldp_linux seems too quiet ...
> > Here is the last captured messages inside ldp_linux in both routers :
> >
> > ....
> > EXIT: ldp_buf_message
> > EXIT: ldp_event
> >
> > After that, nothing.
> >
> > I have tried to figure out the reason but still cannot get it. I am s=
ure
> > that I use the same linux kernel and the same mpls patch in both rout=
ers if
> > it might be the problem.
> >
> > Is it possible that the problem might come from ldp_linux's init mess=
ages,
> > since for determining the host that will be active or passive for sta=
rting
> > the init message you should compare the host addresses (I might be wr=
ong) ?
> > Or simply, the host (interface) that start sending hello messages wil=
l
> > eventually become the active host for init message ?
> >
> > Please give me some clues. In my part, I will start looking your code=
s.
> >
> >
> > regards,
> > ~Venisa Cabrilla
> > _____________________________________________________________________=
____
> > Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.=
com.
> >
> >
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpls-linux-generalOlists.sourceforge.net
> > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>=20
> _______________________________________________
> mpls-linux-general mailing list
> mpls-linux-generalOlists.sourceforge.net
> http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--=20
James R. Leu

Re: [mpls-linux-general] another problem in ldp_linux

[Venisa C. <vca...@ho...>]

Have you downloaded the last ldp-updates using CVS ?
It seems that you are using the CVS version of ldp_linux before Jim sent his 
ldp-updates.

regards,
~Venisa Cabrilla


>From: Vincent Jardin <ja...@en...>
>To: mpl...@li...
>CC: mpl...@li...
>Subject: Re: [mpls-linux-general] another problem in ldp_linux
>Date: Thu, 28 Jun 2001 17:16:23 -0400
>
>I could get quite the same problem, but there is a slight difference. I 
>need
>to configure 10.0.0.6 first.
>
>When it does not work, I get the error:
>   PRT: ldp_state_recv_init: cannot find adj
>on the router 10.0.0.6 when 10.0.0.5 is configured first.
>
>Otherwise, the mpls_in is filled properly when I configure 10.0.0.6 first.
>
>In the case it works (10.0.0.6 is the first router that I configure):
>10.0.0.5 :
>$ cat /proc/net/mpls_in
>40004000 gen 16 0 POP DLV
>40004400 gen 17 0 POP DLV
>40004800 gen 18 0 POP DLV
>but mpls_out, and mpls_fec remain empty.
>Why ?
>
>10.0.0.6:
>$ cat /proc/net/mpls_in
>40004000 gen 16 0 POP DLV
>40004400 gen 17 0 POP DLV
>40004800 gen 18 0 POP DLV
>but mpls_out, and mpls_fec remain empty.
>Why ?
>
>Vincent
>
>Le Mercredi 27 Juin 2001 05:19, Venisa Cabrilla a écrit :
> > Jim,
> >
> > Sorry, another problem ... :-)
> >
> > I know that my problem here will seem somewhat ridiculous but I have 
>spent
> > two and half hours just to find my mistakes ...
> >
> > Please refer to the same configuration I gave you before, but now with
> > mplsadm of course :
> >
> >
> >     lo 10.0.0.6                lo 10.0.0.5
> >      ---------                  ---------
> >
> >     | Router2 |----------------| Router1 |
> >
> >      ---------                  ---------
> >
> >          |eth2  192.168.12.0/24     |eth1
> >
> >     ---------------------------------------
> >
> > In router1:
> > bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > bash# route add 10.0.0.6 gw 192.168.12.2
> > bash# mplsadm -L eth1:0
> >
> > >From inside ldp_linux:
> >
> > prompt> add global 10.0.0.5
> > prompt> add interface eth1
> >
> > In router2:
> > bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > bash# route add 10.0.0.5 gw 192.168.12.1
> > bash# mplsadm -L eth2:0
> >
> > >From inside ldp_linux:
> >
> > prompt> add global 10.0.0.6
> > prompt> add interface eth2
> >
> > I do not know if the order of which we run and configure ldp_linux will
> > certainly determine the result, but in my case, yes !
> >
> > There are two possible cases, we run and configure ldp_linux first in
> > router1 and then in router2, the other, we run and configure ldp_linux 
>in
> > router2 and then router1. What I mean by running and configuring 
>ldp_linux
> > here is that we type add global ... and add interface ...
> >
> > First, we run and configure ldp_linux first in router1(10.0.0.5) and 
>then
> > in router2 (10.0.0.6), ldp_linux works well, I can see that mpls_* are 
>not
> > empty and I can ping the machines. I can see the label mapping process 
>both
> > within ldp_linux or by using ethereal. Everything seems right on its 
>path.
> >
> > Second, we run and configure ldp_linux first in router2(10.0.0.6) and 
>then
> > in router1(10.0.0.1), well here comes the problem, mpls_* rest empty and 
>I
> > can only see hello messages using ethereal, nothing else.
> > There is no label mapping process and ldp_linux seems too quiet ...
> > Here is the last captured messages inside ldp_linux in both routers :
> >
> > ....
> > EXIT: ldp_buf_message
> > EXIT: ldp_event
> >
> > After that, nothing.
> >
> > I have tried to figure out the reason but still cannot get it. I am sure
> > that I use the same linux kernel and the same mpls patch in both routers 
>if
> > it might be the problem.
> >
> > Is it possible that the problem might come from ldp_linux's init 
>messages,
> > since for determining the host that will be active or passive for 
>starting
> > the init message you should compare the host addresses (I might be 
>wrong) ?
> > Or simply, the host (interface) that start sending hello messages will
> > eventually become the active host for init message ?
> >
> > Please give me some clues. In my part, I will start looking your codes.
> >
> >
> > regards,
> > ~Venisa Cabrilla
> > 
>_________________________________________________________________________
> > Get Your Private, Free E-mail from MSN Hotmail at 
>http://www.hotmail.com.
> >
> >
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
>_______________________________________________
>mpls-linux-general mailing list
>mpl...@li...
>http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Re: [mpls-linux-general] another problem in ldp_linux

[Vincent J. <ja...@en...>]

I could get quite the same problem, but there is a slight difference. I need 
to configure 10.0.0.6 first.

When it does not work, I get the error:
  PRT: ldp_state_recv_init: cannot find adj
on the router 10.0.0.6 when 10.0.0.5 is configured first.

Otherwise, the mpls_in is filled properly when I configure 10.0.0.6 first.

In the case it works (10.0.0.6 is the first router that I configure):
10.0.0.5 :
$ cat /proc/net/mpls_in  
40004000 gen 16 0 POP DLV 
40004400 gen 17 0 POP DLV 
40004800 gen 18 0 POP DLV
but mpls_out, and mpls_fec remain empty.
Why ?

10.0.0.6:
$ cat /proc/net/mpls_in  
40004000 gen 16 0 POP DLV 
40004400 gen 17 0 POP DLV 
40004800 gen 18 0 POP DLV 
but mpls_out, and mpls_fec remain empty.
Why ?

Vincent

Le Mercredi 27 Juin 2001 05:19, Venisa Cabrilla a écrit :
> Jim,
>
> Sorry, another problem ... :-)
>
> I know that my problem here will seem somewhat ridiculous but I have spent
> two and half hours just to find my mistakes ...
>
> Please refer to the same configuration I gave you before, but now with
> mplsadm of course :
>
>
>     lo 10.0.0.6                lo 10.0.0.5
>      ---------                  ---------
>
>     | Router2 |----------------| Router1 |
>
>      ---------                  ---------
>
>          |eth2  192.168.12.0/24     |eth1
>
>     ---------------------------------------
>
> In router1:
> bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> bash# route add 10.0.0.6 gw 192.168.12.2
> bash# mplsadm -L eth1:0
>
> >From inside ldp_linux:
>
> prompt> add global 10.0.0.5
> prompt> add interface eth1
>
> In router2:
> bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> bash# route add 10.0.0.5 gw 192.168.12.1
> bash# mplsadm -L eth2:0
>
> >From inside ldp_linux:
>
> prompt> add global 10.0.0.6
> prompt> add interface eth2
>
> I do not know if the order of which we run and configure ldp_linux will
> certainly determine the result, but in my case, yes !
>
> There are two possible cases, we run and configure ldp_linux first in
> router1 and then in router2, the other, we run and configure ldp_linux in
> router2 and then router1. What I mean by running and configuring ldp_linux
> here is that we type add global ... and add interface ...
>
> First, we run and configure ldp_linux first in router1(10.0.0.5) and then
> in router2 (10.0.0.6), ldp_linux works well, I can see that mpls_* are not
> empty and I can ping the machines. I can see the label mapping process both
> within ldp_linux or by using ethereal. Everything seems right on its path.
>
> Second, we run and configure ldp_linux first in router2(10.0.0.6) and then
> in router1(10.0.0.1), well here comes the problem, mpls_* rest empty and I
> can only see hello messages using ethereal, nothing else.
> There is no label mapping process and ldp_linux seems too quiet ...
> Here is the last captured messages inside ldp_linux in both routers :
>
> ....
> EXIT: ldp_buf_message
> EXIT: ldp_event
>
> After that, nothing.
>
> I have tried to figure out the reason but still cannot get it. I am sure
> that I use the same linux kernel and the same mpls patch in both routers if
> it might be the problem.
>
> Is it possible that the problem might come from ldp_linux's init messages,
> since for determining the host that will be active or passive for starting
> the init message you should compare the host addresses (I might be wrong) ?
> Or simply, the host (interface) that start sending hello messages will
> eventually become the active host for init message ?
>
> Please give me some clues. In my part, I will start looking your codes.
>
>
> regards,
> ~Venisa Cabrilla
> _________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
>
>
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

Re: [mpls-linux-general] LDP TIMERS

[Venisa C. <vca...@ho...>]

>From: "nomit kalidhar" <nom...@in...>
>Subject: [mpls-linux-general] LDP TIMERS
>Date: Wed, 27 Jun 2001 15:12:23 +0530
>
>Hello,
>
>I had a small doubt in LDP regarding timers.
>
>A LSR maintains a Hold timer with each Hello adjacency which it restarts
>when it receives a hello message.
>  When a session is established does it still have to exchange Hello
>packets ? along with keep alive message ( if nothing is there to
>transmit) ?

Yes, if there is no PDU to transmit, we use the keep alive message to 
restart the keep alive timer.

>If it keeps on sending the hello message and the keep alive message,
>then what happens if the hold timer gets over but keep alive is still
>not over, is the connection terminated ?

Yes, if the hold timer expires, the LSR deletes the hello adjacency, then 
terminates the LDP session by sending a notification message and closes the 
transport connection ( you can refer to RFC 3036 session 2.5.5. Maintaining 
Hello Adjacencies :-) ).


>TIA
>
>Nomit

regards,
~Venisa Cabrilla
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

[mpls-linux-general] Re: misunderstanding

[Venisa C. <vca...@ho...>]

>From: "Giovanna Piantanida (ERI)" <Gio...@er...>
>To: "'Venisa Cabrilla'" <vca...@ho...>
>Subject: misunderstanding
>Date: Thu, 28 Jun 2001 10:48:44 +0200
>
>It is of course my fault, I realise now it wasn't completely clear.
>
>I meant that ldp_zebra was the one not working well!!!
>
>I know ldp_linux works in the way you noticed:
>-release 0.060 only put mpls_in labels
>-CVS release fixed Unsolicited Mode but not the labelspace

Exactly.

>but I never tried to couple it with zebra, on the other hand I
>got the impression, from your email, that you did it and you
>found a good interaction, is that it?
>regards
>Giovanna
>
>

Yes, so far they have a good interaction. Indeed, I use zebra just to 
provide the routing information for ldp_linux so that I do not need to 
explicitly add the routes in each routers I have. Since I have some routers 
in my network, it is really inconvenient if I have to add manually the 
routes by route commands.
So according to your experiences, ldp_zebra is not working yet ?

>
>-----Original Message-----
>From: Venisa Cabrilla [mailto:vca...@ho...]
>Sent: giovedì 28 giugno 2001 10.02
>To: Gio...@er...
>Subject: RE: [mpls-linux-general] a bug ?
>
>
>Please see my comments inline.
>
> >From: "Giovanna Piantanida (ERI)" <Gio...@er...>
> >To: "'Venisa Cabrilla'" <vca...@ho...>
> >Subject: RE: [mpls-linux-general] a bug ?
> >Date: Thu, 28 Jun 2001 09:40:46 +0200
> >
> >You mean that you ran ldp_linux with the zebra daemon? Because up-to-now
> >I only tried ldp_linux alone and ldp_zebra with zebra, and I observed 
>that
> >it sends the messages,
> >the zebra databases are acknowledged about the label mappings, but labels
> >are
> >never written in mpls_in, mpls_out etc..etc.. so, traffic is not labeled
> >with mpls.
>
>Yes, I ran ldp_linux with zebra daemon (also ospfd and bgpd) and it worked
>very well as long as I apply the "right" order. The mpls_* files were not
>empty and I could see the MPLS encapsulation using ethereal when I pinged
>the machines.
>I had the same problems as you stated (see my posting "problem in
>ldp_linux"), but that just before I downloaded the CVS version of
>ldp-portable. Only the mpls_in was not empty.
>I believe that you have to use the CVS version of ldp-portable since Jim 
>has
>done some fixes.
>That's why I ask Jim what is the main difference between ldp_linux and
>ldp_zebra if I can (I did) run ldp_linux with zebra.
>Since you had tested ldp_zebra with zebra, do you see the difference ?
>
> >Moreover, I believe that it has to be something in the new files which
> >causes your bug,
> >because I never found out something like that, and I used to run 
>ldp_linux
> >on several routers
> >in different order.
>
>Yes, that's what Jim has explained.
>
> >regards,
> >Giovanna Piantanida
> >
>
>regards,
>~Venisa Cabrilla
>_________________________________________________________________________
>Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Re: [mpls-linux-general] a bug ?

[Venisa C. <vca...@ho...>]

Jim,

Despite that little bug, I have tested ldp_linux running with zebra (in the 
"right" order) and so far it works very well.
Now, I will test ldp_zebra to see its differences from ldp_linux ...

regards,
~Venisa Cabrilla



>From: "James R. Leu" <jl...@mi...>
>Reply-To: jl...@mi...
>To: Venisa Cabrilla <vca...@ho...>
>CC: mpl...@li...
>Subject: Re: [mpls-linux-general] a bug ?
>Date: Wed, 27 Jun 2001 11:45:54 -0500
>
>On Wed, Jun 27, 2001 at 11:36:29AM -0000, Venisa Cabrilla wrote:
> > Jim,
> >
> > I use the third router(lo 10.0.0.7) directly connected to router2 (lo
> > 10.0.0.6) to support my argumentation here related to my last posting
> > ("another problem in ldp_linux").
> >
> > It seems that ldp_linux works fine if I run and configure it from 
>interface
> > (lo) whose address is smaller. In my case of using three routers, 
>ldp_linux
> > works well, if and only if I start configuring from router1(lo 10.0.0.5) 
>to
> > router2(lo 10.0.0.6) and then from router2(lo 10.0.0.6) to router3(lo
> > 10.0.0.7).
> >
> > Am I right ? Do we always have to do like that ?
>
>There must be a bug with ldp_linux.  LDP determins who will be "active"
>based on the LSRID (router ids).  The one with the higher address is 
>active.
>When you start them in the "wrong" order the LSR that is active hasn't
>had time to send a hello yet.  Thus the 'passive' LSR doen't know of it's
>existence and when the 'active' LSR tries to setup the TCP session it 
>fails.
>
>What should happen next is that the 'active' LSR goes into backoff state 
>for
>about 10 seconds, then tries again.  If your not seeing this then something
>is broken.
>
>(BTW I had this tested and working before my recent TCP over haul it must
>be a bug in there)
>
>I'll take a look at it tonight and see if I can fix it.
>
>Thanks for the bug report!
>
>Jim
>--
>James R. Leu
>
>_______________________________________________
>mpls-linux-general mailing list
>mpl...@li...
>http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Re: [mpls-linux-general] about tunneling.

[Srinivas R. <ls...@tr...>]

thank you jim,


regards
lsreddy

On Tuesday 26 June 2001 21:09, you wrote:
> On Tue, Jun 26, 2001 at 04:21:43PM +0530, Srinivas Reddy wrote:
> > hello all ,
> >
> > 	i am getting confussion about tunneling. practically in which case we
> > get tunneling. when we get tunneling we have to send Targetted Hello
> > only(please correct me if i am wrong). Can some body guide me in which
> > cases we will create tunnel especially in the case if we use BGP .
>
> I'm not sure if I understand your question but here is an attempt to
> answer it :-)
>
> BTW I hate the term "tunnel" (as in an MPLS tunnel) but since you've used
> it, I will use the term and explain what I mean by it.
>
>
> If we have a signalling protocol set up an LSP to a remote router AND
> the signalling protocol associats the remote router ID with the LSP,
> then we have an instance where "tunneling" is feasible.  In this case
> "tunneling" means that I will treat the remote router as a directly
> connected peer (via the LSP) for the purpose of another protocol
> (signalling OR routing).
>
> Stepping back.  Signalling protocol that setup LSPs and associate router
> IDs are:
>
> RSVP-TE (user is responsible for assigning a "destination" address,
> in this case the destination address should be the remote router ID)
>
> CR-LDP (the FEC could be the remote router ID /32 or the user is
> responsible for assigning a "destination" address, same as for RSVP-TE)
>
> LDP (the FEC is the remote router ID and is normally sent via unsolicited
> mode (ala Juniper))
>
> In any of these cases the result will be an LSP which can be associated
> with the remote router ID (how the association is made varies).
>
> Once that assocaition is made then the remote router is can be considered
> as directly connect (as you would say, a "tunnel" has been created to the
> remote router).
>
> Now a second level of protocols can use this "tunnel".  Examples are:
>
> LDP (requires "tunnels" in both directions and then uses targeted hellos
> to establish an adj.  The resulting session needs to be associated
> with a "tunnel" for the label mappings to be used)
>
> RSVP-TE (the "tunnel" is treated as one hop in an explicit hop list)
>
> BGP (requires "tunnels" in both directions and each side must advertise
> their router ID as the next hop for each NRLI thus BGP can resolve the
> nexthop via the "tunnel")
>
> One note about the second level signalling protocols:
>
> The resulting label allocation made by second level signalling protocols
> are only known by the end point of the LSP.  Thus packet traversing the
> LSPs created by second level signalling protocols have 2 labels on them. 
> The top label is for the "tunnel" LSP and the bottom label is for the
> LSP created by the second level signalling protocol.
>
> I hope this answers your question :-)
>
> Jim
>
> PS In the above example you can replace router ID with "network wide unique
> address that is 'known' to be assocaited with the remote router and is
> still valid even when re-routing occurs due to link failures".  :-)
>
> > thanks in advance
> > regards
> > lsr
> >
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

Re: [mpls-linux-general] a bug ?

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

On Wed, Jun 27, 2001 at 11:36:29AM -0000, Venisa Cabrilla wrote:
> Jim,
> 
> I use the third router(lo 10.0.0.7) directly connected to router2 (lo 
> 10.0.0.6) to support my argumentation here related to my last posting 
> ("another problem in ldp_linux").
> 
> It seems that ldp_linux works fine if I run and configure it from interface 
> (lo) whose address is smaller. In my case of using three routers, ldp_linux 
> works well, if and only if I start configuring from router1(lo 10.0.0.5) to 
> router2(lo 10.0.0.6) and then from router2(lo 10.0.0.6) to router3(lo 
> 10.0.0.7).
> 
> Am I right ? Do we always have to do like that ?

There must be a bug with ldp_linux.  LDP determins who will be "active"
based on the LSRID (router ids).  The one with the higher address is active.
When you start them in the "wrong" order the LSR that is active hasn't
had time to send a hello yet.  Thus the 'passive' LSR doen't know of it's
existence and when the 'active' LSR tries to setup the TCP session it fails.

What should happen next is that the 'active' LSR goes into backoff state for
about 10 seconds, then tries again.  If your not seeing this then something
is broken.

(BTW I had this tested and working before my recent TCP over haul it must
be a bug in there)

I'll take a look at it tonight and see if I can fix it.

Thanks for the bug report!

Jim
-- 
James R. Leu

[mpls-linux-general] a bug ?

[Venisa C. <vca...@ho...>]

Jim,

I use the third router(lo 10.0.0.7) directly connected to router2 (lo 
10.0.0.6) to support my argumentation here related to my last posting 
("another problem in ldp_linux").

It seems that ldp_linux works fine if I run and configure it from interface 
(lo) whose address is smaller. In my case of using three routers, ldp_linux 
works well, if and only if I start configuring from router1(lo 10.0.0.5) to 
router2(lo 10.0.0.6) and then from router2(lo 10.0.0.6) to router3(lo 
10.0.0.7).

Am I right ? Do we always have to do like that ?

regards,
~Venisa Cabrilla
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

[mpls-linux-general] LDP TIMERS

[nomit k. <nom...@in...>]

Hello,

I had a small doubt in LDP regarding timers.

A LSR maintains a Hold timer with each Hello adjacency which it restarts
when it receives a hello message.
 When a session is established does it still have to exchange Hello
packets ? along with keep alive message ( if nothing is there to
transmit) ?

If it keeps on sending the hello message and the keep alive message,
then what happens if the hold timer gets over but keep alive is still
not over, is the connection terminated ?

TIA

Nomit

[mpls-linux-general] another problem in ldp_linux

[Venisa C. <vca...@ho...>]

Jim,

Sorry, another problem ... :-)

I know that my problem here will seem somewhat ridiculous but I have spent 
two and half hours just to find my mistakes ...

Please refer to the same configuration I gave you before, but now with 
mplsadm of course :


    lo 10.0.0.6                lo 10.0.0.5
     ---------                  ---------
    | Router2 |----------------| Router1 |
     ---------                  ---------
         |eth2  192.168.12.0/24     |eth1
    ---------------------------------------

In router1:
bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
bash# route add 10.0.0.6 gw 192.168.12.2
bash# mplsadm -L eth1:0

From inside ldp_linux:
prompt> add global 10.0.0.5
prompt> add interface eth1

In router2:
bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
bash# route add 10.0.0.5 gw 192.168.12.1
bash# mplsadm -L eth2:0

From inside ldp_linux:
prompt> add global 10.0.0.6
prompt> add interface eth2

I do not know if the order of which we run and configure ldp_linux will 
certainly determine the result, but in my case, yes !

There are two possible cases, we run and configure ldp_linux first in 
router1 and then in router2, the other, we run and configure ldp_linux in 
router2 and then router1. What I mean by running and configuring ldp_linux 
here is that we type add global ... and add interface ...

First, we run and configure ldp_linux first in router1(10.0.0.5) and then in 
router2 (10.0.0.6), ldp_linux works well, I can see that mpls_* are not 
empty and I can ping the machines. I can see the label mapping process both 
within ldp_linux or by using ethereal. Everything seems right on its path.

Second, we run and configure ldp_linux first in router2(10.0.0.6) and then 
in router1(10.0.0.1), well here comes the problem, mpls_* rest empty and I 
can only see hello messages using ethereal, nothing else.
There is no label mapping process and ldp_linux seems too quiet ...
Here is the last captured messages inside ldp_linux in both routers :

...
EXIT: ldp_buf_message
EXIT: ldp_event

After that, nothing.

I have tried to figure out the reason but still cannot get it. I am sure 
that I use the same linux kernel and the same mpls patch in both routers if 
it might be the problem.

Is it possible that the problem might come from ldp_linux's init messages, 
since for determining the host that will be active or passive for starting 
the init message you should compare the host addresses (I might be wrong) ? 
Or simply, the host (interface) that start sending hello messages will 
eventually become the active host for init message ?

Please give me some clues. In my part, I will start looking your codes.


regards,
~Venisa Cabrilla
_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Re: [mpls-linux-general] ldp-portable updates

[Venisa C. <vca...@ho...>]

Jim,

ldp_zebra ? Uhmmm, infact, I am still confused %-) about the real difference 
between ldp_linux and ldp_zebra (I have not tested ldp_zebra yet).

Suppose that I can run ldp_linux with zebra, that is zebra will provide the 
routing information to ldp_linux, in this case what makes ldp_zebra so 
different since for making it work we also have to run zebra before (that's 
at least what I can understand from your README).

Or maybe in ldp_zebra, the relation between ldp_zebra and zebra is just like 
the relation between ospfd or bgpd with zebra ?

BTW I am trying to run ldp_linux with zebra, but I still have some problems 
with it, I'll hope I still can overcome it ...


regards,
~Venisa Cabrilla



>From: "James R. Leu" <jl...@mi...>
>Reply-To: jl...@mi...
>To: Venisa Cabrilla <vca...@ho...>
>CC: mpl...@li...
>Subject: Re: [mpls-linux-general] ldp-portable updates
>Date: Tue, 26 Jun 2001 10:44:42 -0500
>
>On Tue, Jun 26, 2001 at 03:24:39PM -0000, Venisa Cabrilla wrote:
> > That's it ! It works now :-D Thanks Jim ...
>
>Ahhhh.  Another satisfied ldp-portable user ;-)
>
>For my next trick we'll get it working with ldp_zebra ...
>
>This weill require a bit more code from me ....
>
>Jim
>
> >
> > regards,
> > ~Venisa Cabrilla
> >
> >
> > >From: "James R. Leu" <jl...@mi...>
> > >Reply-To: jl...@mi...
> > >To: Venisa Cabrilla <vca...@ho...>
> > >CC: mpl...@li...
> > >Subject: Re: [mpls-linux-general] ldp-portable updates
> > >Date: Tue, 26 Jun 2001 09:57:31 -0500
> > >
> > >On Tue, Jun 26, 2001 at 02:28:11PM -0000, Venisa Cabrilla wrote:
> > > > Hi,
> > > >
> > > > Sorry for the late response, I have to go downstair and upstair just 
>to
> > >copy
> > > > the contents of /proc/net/mpls_*, (my lab and my office are not in 
>the
> > >same
> > > > room) but I have done as fast as I can ... :-)
> > >
> > >Late response!  I'm used to latency on the order of days (or weeks ;-)
> > >Your latency is spetacular! :-)
> > >
> > > > First, I have to inform you that since I use my computers as 
>routers, I
> > >do
> > > > certainly have at least three ethernet cards in my routers (that is, 
>the
> > > > topology that I gave you is far from complete, which will explain 
>the
> > > > important contents of file mpls_in) but I used only the commands 
>that I
> > >have
> > > > described in my last email.
> > >
> > >This is cool!  ldp-portable is actually working for someone beside me!
> > >
> > > > In router1:
> > > > mpls_in:
> > > > 40004000 gen 16 0 POP DLV
> > > > 40004400 gen 17 0 POP DLV
> > > > 40004800 gen 18 0 POP DLV
> > > > 40004c00 gen 19 0 POP DLV
> > > > 40005000 gen 20 0 POP DLV
> > > > 40005400 gen 21 0 POP DLV
> > > > 40005800 gen 22 0 POP DLV
> > > > mpls_out:
> > > > 40004003 PUSH(gen 16) SET (eth1)
> > > > mpls_fec:
> > > > 40004003 10.0.0.6/32
> > >
> > >Ahhhh.  Here is the culprite.
> > >
> > > > mpls_labelspace:
> > > > lo 0
> > >
> > >ldp_linux doesn't set the label space for you.  You currently need to
> > >execute
> > >the command 'mplsadm -L eth1:0' for each of the LDP speaking interfaces
> > >before
> > >starting up ldp_linux.
> > >
> > >I have a patch from someone at Ericsson that fixes this.
> > >
> > >Other then the labelspace problem the rest look just fine :-)
> > >
> > >Let me know if this fixes it.
> > >
> > >BTW clues to this being the problem would have shown up in the MPLS 
>kernel
> > >debugging when you tried to ping.
> > >
> > >On Router1:
> > >mplsadm -d
> > >ping 10.0.0.6
> > >dmesg
> > >
> > >You should see that the packet are being sent via mpls_output()
> > >
> > >On Router2:
> > >ping 10.0.0.5
> > >
> > >On Router1:
> > >dmesg
> > >
> > >You should see info about packets arriving on interfaces without a
> > >labelspace
> > >
> > >Jim
> > >
> > > > In router2:
> > > > mpls_in:
> > > > 40004000 gen 16 0 POP DLV
> > > > 40004400 gen 17 0 POP DLV
> > > > 40004800 gen 18 0 POP DLV
> > > > 40004c00 gen 19 0 POP DLV
> > > > 40005000 gen 20 0 POP DLV
> > > > 40005400 gen 21 0 POP DLV
> > > > 40005800 gen 22 0 POP DLV
> > > > mpls_out:
> > > > 40004004 PUSH(gen 16) SET (eth2)
> > > > mpls_fec:
> > > > 40004004 10.0.0.5/32
> > > > mpls_labelspace:
> > > > lo 0
> > > >
> > > > There is nothing in mpls_tunnel, which is normal.
> > > >
> > > >
> > > > regards,
> > > > ~Venisa Cabrilla
> > > >
> > > >
> > > >
> > > > >From: "James R. Leu" <jl...@mi...>
> > > > >Reply-To: jl...@mi...
> > > > >To: Venisa Cabrilla <vca...@ho...>
> > > > >CC: mpl...@li...
> > > > >Subject: Re: [mpls-linux-general] ldp-portable updates
> > > > >Date: Tue, 26 Jun 2001 08:45:55 -0500
> > > > >
> > > > >On Tue, Jun 26, 2001 at 01:31:11PM -0000, Venisa Cabrilla wrote:
> > > > > > Hi James,
> > > > > >
> > > > > > Thanks for your ldp-portable updates, now I can see that 
>mpls_in,
> > > > >mpls_out,
> > > > > > and mpls_fec are not empty.
> > > > >
> > > > >Hello,
> > > > >
> > > > >I think I know the problem.  It has to do with the fact that LDP is
> > >setting
> > > > >up
> > > > >LSPs for too many entries in the route table.  Send me the contents 
>of
> > > > >/proc/net/mpls_* and we can verify this.
> > > > >
> > > > >My default policy of "send labels for everything" may not be valid 
>in
> > >all
> > > > >cases.  This will be simple enough to change I just need to figure 
>out
> > >what
> > > > >the
> > > > >best default policy should be :-(
> > > > >
> > > > >I'll try to duplicate this setup here and see if I can figure it 
>out as
> > > > >well.
> > > > >
> > > > >Jim
> > > > >
> > > > > >
> > > > > > I use the simple configuration in file README.sample :
> > > > > >
> > > > > >      lo 10.0.0.6                lo 10.0.0.5
> > > > > >       ---------                  ---------
> > > > > >      | Router2 |----------------| Router1 |
> > > > > >       ---------                  ---------
> > > > > >           |eth2  192.168.12.0/24     |eth1
> > > > > >      ---------------------------------------
> > > > > >
> > > > > > In router1:
> > > > > > bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > > > > > bash# route add 10.0.0.6 gw 192.168.12.2
> > > > > >
> > > > > > >From inside ldp_linux:
> > > > > > prompt> add global 10.0.0.5
> > > > > > prompt> add interface eth1
> > > > > >
> > > > > > In router2:
> > > > > > bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > > > > > bash# route add 10.0.0.5 gw 192.168.12.1
> > > > > >
> > > > > > >From inside ldp_linux:
> > > > > > prompt> add global 10.0.0.6
> > > > > > prompt> add interface eth2
> > > > > >
> > > > > > After that I can see that mpls_in, mpls_out, and mpls_fec in 
>each
> > >router
> > > > >are
> > > > > > not empty.
> > > > > >
> > > > > > Unfortunately, I cannot ping anymore router1 from router2 and
> > >router2
> > > > >from
> > > > > > router1. Before I enter ldp_linux, I can ping 10.0.0.5 from 
>router1
> > >and
> > > > >ping
> > > > > > 10.0.0.6 from router2. However, I can see the MPLS encapsulation
> > >using
> > > > > > ethereal for ping request messages but there is no ping reply
> > >message.
> > > > > >
> > > > > > Would you mind verify my configuration above and tell me what 
>might
> > >be
> > > > >my
> > > > > > mistakes ?
> > > > > >
> > > > > >
> > > > > > PS: I have to explicitly add the routes using route commands 
>this
> > >time,
> > > > > > because if I use zebra, I cannot use ethereal (ethereal hangs).
> > > > > > I still do not know the reason, it might be due to TCP/UDP
> > >conflicts.
> > > > > >
> > > > > >
> > > > > > regards,
> > > > > > ~Venisa Cabrilla
> > > > > >
> > > > > >
> > > > > >
> > > > > > >From: "James R. Leu" <jl...@mi...>
> > > > > > >Reply-To: jl...@mi...
> > > > > > >To: mpl...@li...
> > > > > > >Subject: [mpls-linux-general] ldp-portable updates
> > > > > > >Date: Sun, 24 Jun 2001 20:23:13 -0500
> > > > > > >
> > > > > > >Sorry to those who have been trying to use ldp-portable from 
>CVS.
> > > > > > >I finnally have gotten around to fixing and testing 
>ldp-portable.
> > > > > > >It's almost back to where it was before I started mucking 
>around
> > >with
> > > > >PDU
> > > > > > >processing.  (interoping with cisco pointed out the need to 
>handle
> > > > >multiple
> > > > > > >PDUs per TCP segment and handle one PDU spread across multiple 
>TCP
> > > > > > >segments)
> > > > > > >
> > > > > > >If you tried ldp-portable in the last 3 weeks and it was 
>broken,
> > >try
> > > > > > >updating your CVS client and giving it another go.
> > > > > > >
> > > > > > >Sorry for the long delay.
> > > > > > >Jim
> > > > > > >--
> > > > > > >James R. Leu
> > > > > > >
> > > > > > >_______________________________________________
> > > > > > >mpls-linux-general mailing list
> > > > > > >mpl...@li...
> > > > > > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > > >
> > > > > >
> > > >
> > > 
> >_________________________________________________________________________
> > > > > > Get Your Private, Free E-mail from MSN Hotmail at
> > > > >http://www.hotmail.com.
> > > > > >
> > > > > >
> > > > > > _______________________________________________
> > > > > > mpls-linux-general mailing list
> > > > > > mpl...@li...
> > > > > > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > >
> > > > >--
> > > > >James R. Leu
> > > > >
> > > > >_______________________________________________
> > > > >mpls-linux-general mailing list
> > > > >mpl...@li...
> > > > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > >
> > > >
> > 
> >_________________________________________________________________________
> > > > Get Your Private, Free E-mail from MSN Hotmail at
> > >http://www.hotmail.com.
> > >
> > >--
> > >James R. Leu
> >
> > 
>_________________________________________________________________________
> > Get Your Private, Free E-mail from MSN Hotmail at 
>http://www.hotmail.com.
> >
> >
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
>--
>James R. Leu
>
>_______________________________________________
>mpls-linux-general mailing list
>mpl...@li...
>http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Re: [mpls-linux-general] ldp-portable updates

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

On Tue, Jun 26, 2001 at 03:24:39PM -0000, Venisa Cabrilla wrote:
> That's it ! It works now :-D Thanks Jim ...

Ahhhh.  Another satisfied ldp-portable user ;-)

For my next trick we'll get it working with ldp_zebra ...

This weill require a bit more code from me ....

Jim

> 
> regards,
> ~Venisa Cabrilla
> 
> 
> >From: "James R. Leu" <jl...@mi...>
> >Reply-To: jl...@mi...
> >To: Venisa Cabrilla <vca...@ho...>
> >CC: mpl...@li...
> >Subject: Re: [mpls-linux-general] ldp-portable updates
> >Date: Tue, 26 Jun 2001 09:57:31 -0500
> >
> >On Tue, Jun 26, 2001 at 02:28:11PM -0000, Venisa Cabrilla wrote:
> > > Hi,
> > >
> > > Sorry for the late response, I have to go downstair and upstair just to 
> >copy
> > > the contents of /proc/net/mpls_*, (my lab and my office are not in the 
> >same
> > > room) but I have done as fast as I can ... :-)
> >
> >Late response!  I'm used to latency on the order of days (or weeks ;-)
> >Your latency is spetacular! :-)
> >
> > > First, I have to inform you that since I use my computers as routers, I 
> >do
> > > certainly have at least three ethernet cards in my routers (that is, the
> > > topology that I gave you is far from complete, which will explain the
> > > important contents of file mpls_in) but I used only the commands that I 
> >have
> > > described in my last email.
> >
> >This is cool!  ldp-portable is actually working for someone beside me!
> >
> > > In router1:
> > > mpls_in:
> > > 40004000 gen 16 0 POP DLV
> > > 40004400 gen 17 0 POP DLV
> > > 40004800 gen 18 0 POP DLV
> > > 40004c00 gen 19 0 POP DLV
> > > 40005000 gen 20 0 POP DLV
> > > 40005400 gen 21 0 POP DLV
> > > 40005800 gen 22 0 POP DLV
> > > mpls_out:
> > > 40004003 PUSH(gen 16) SET (eth1)
> > > mpls_fec:
> > > 40004003 10.0.0.6/32
> >
> >Ahhhh.  Here is the culprite.
> >
> > > mpls_labelspace:
> > > lo 0
> >
> >ldp_linux doesn't set the label space for you.  You currently need to 
> >execute
> >the command 'mplsadm -L eth1:0' for each of the LDP speaking interfaces 
> >before
> >starting up ldp_linux.
> >
> >I have a patch from someone at Ericsson that fixes this.
> >
> >Other then the labelspace problem the rest look just fine :-)
> >
> >Let me know if this fixes it.
> >
> >BTW clues to this being the problem would have shown up in the MPLS kernel
> >debugging when you tried to ping.
> >
> >On Router1:
> >mplsadm -d
> >ping 10.0.0.6
> >dmesg
> >
> >You should see that the packet are being sent via mpls_output()
> >
> >On Router2:
> >ping 10.0.0.5
> >
> >On Router1:
> >dmesg
> >
> >You should see info about packets arriving on interfaces without a 
> >labelspace
> >
> >Jim
> >
> > > In router2:
> > > mpls_in:
> > > 40004000 gen 16 0 POP DLV
> > > 40004400 gen 17 0 POP DLV
> > > 40004800 gen 18 0 POP DLV
> > > 40004c00 gen 19 0 POP DLV
> > > 40005000 gen 20 0 POP DLV
> > > 40005400 gen 21 0 POP DLV
> > > 40005800 gen 22 0 POP DLV
> > > mpls_out:
> > > 40004004 PUSH(gen 16) SET (eth2)
> > > mpls_fec:
> > > 40004004 10.0.0.5/32
> > > mpls_labelspace:
> > > lo 0
> > >
> > > There is nothing in mpls_tunnel, which is normal.
> > >
> > >
> > > regards,
> > > ~Venisa Cabrilla
> > >
> > >
> > >
> > > >From: "James R. Leu" <jl...@mi...>
> > > >Reply-To: jl...@mi...
> > > >To: Venisa Cabrilla <vca...@ho...>
> > > >CC: mpl...@li...
> > > >Subject: Re: [mpls-linux-general] ldp-portable updates
> > > >Date: Tue, 26 Jun 2001 08:45:55 -0500
> > > >
> > > >On Tue, Jun 26, 2001 at 01:31:11PM -0000, Venisa Cabrilla wrote:
> > > > > Hi James,
> > > > >
> > > > > Thanks for your ldp-portable updates, now I can see that mpls_in,
> > > >mpls_out,
> > > > > and mpls_fec are not empty.
> > > >
> > > >Hello,
> > > >
> > > >I think I know the problem.  It has to do with the fact that LDP is 
> >setting
> > > >up
> > > >LSPs for too many entries in the route table.  Send me the contents of
> > > >/proc/net/mpls_* and we can verify this.
> > > >
> > > >My default policy of "send labels for everything" may not be valid in 
> >all
> > > >cases.  This will be simple enough to change I just need to figure out 
> >what
> > > >the
> > > >best default policy should be :-(
> > > >
> > > >I'll try to duplicate this setup here and see if I can figure it out as
> > > >well.
> > > >
> > > >Jim
> > > >
> > > > >
> > > > > I use the simple configuration in file README.sample :
> > > > >
> > > > >      lo 10.0.0.6                lo 10.0.0.5
> > > > >       ---------                  ---------
> > > > >      | Router2 |----------------| Router1 |
> > > > >       ---------                  ---------
> > > > >           |eth2  192.168.12.0/24     |eth1
> > > > >      ---------------------------------------
> > > > >
> > > > > In router1:
> > > > > bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > > > > bash# route add 10.0.0.6 gw 192.168.12.2
> > > > >
> > > > > >From inside ldp_linux:
> > > > > prompt> add global 10.0.0.5
> > > > > prompt> add interface eth1
> > > > >
> > > > > In router2:
> > > > > bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > > > > bash# route add 10.0.0.5 gw 192.168.12.1
> > > > >
> > > > > >From inside ldp_linux:
> > > > > prompt> add global 10.0.0.6
> > > > > prompt> add interface eth2
> > > > >
> > > > > After that I can see that mpls_in, mpls_out, and mpls_fec in each 
> >router
> > > >are
> > > > > not empty.
> > > > >
> > > > > Unfortunately, I cannot ping anymore router1 from router2 and 
> >router2
> > > >from
> > > > > router1. Before I enter ldp_linux, I can ping 10.0.0.5 from router1 
> >and
> > > >ping
> > > > > 10.0.0.6 from router2. However, I can see the MPLS encapsulation 
> >using
> > > > > ethereal for ping request messages but there is no ping reply 
> >message.
> > > > >
> > > > > Would you mind verify my configuration above and tell me what might 
> >be
> > > >my
> > > > > mistakes ?
> > > > >
> > > > >
> > > > > PS: I have to explicitly add the routes using route commands this 
> >time,
> > > > > because if I use zebra, I cannot use ethereal (ethereal hangs).
> > > > > I still do not know the reason, it might be due to TCP/UDP 
> >conflicts.
> > > > >
> > > > >
> > > > > regards,
> > > > > ~Venisa Cabrilla
> > > > >
> > > > >
> > > > >
> > > > > >From: "James R. Leu" <jl...@mi...>
> > > > > >Reply-To: jl...@mi...
> > > > > >To: mpl...@li...
> > > > > >Subject: [mpls-linux-general] ldp-portable updates
> > > > > >Date: Sun, 24 Jun 2001 20:23:13 -0500
> > > > > >
> > > > > >Sorry to those who have been trying to use ldp-portable from CVS.
> > > > > >I finnally have gotten around to fixing and testing ldp-portable.
> > > > > >It's almost back to where it was before I started mucking around 
> >with
> > > >PDU
> > > > > >processing.  (interoping with cisco pointed out the need to handle
> > > >multiple
> > > > > >PDUs per TCP segment and handle one PDU spread across multiple TCP
> > > > > >segments)
> > > > > >
> > > > > >If you tried ldp-portable in the last 3 weeks and it was broken, 
> >try
> > > > > >updating your CVS client and giving it another go.
> > > > > >
> > > > > >Sorry for the long delay.
> > > > > >Jim
> > > > > >--
> > > > > >James R. Leu
> > > > > >
> > > > > >_______________________________________________
> > > > > >mpls-linux-general mailing list
> > > > > >mpl...@li...
> > > > > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > >
> > > > >
> > > 
> > >_________________________________________________________________________
> > > > > Get Your Private, Free E-mail from MSN Hotmail at
> > > >http://www.hotmail.com.
> > > > >
> > > > >
> > > > > _______________________________________________
> > > > > mpls-linux-general mailing list
> > > > > mpl...@li...
> > > > > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > >
> > > >--
> > > >James R. Leu
> > > >
> > > >_______________________________________________
> > > >mpls-linux-general mailing list
> > > >mpl...@li...
> > > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > >
> > > 
> >_________________________________________________________________________
> > > Get Your Private, Free E-mail from MSN Hotmail at 
> >http://www.hotmail.com.
> >
> >--
> >James R. Leu
> 
> _________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

Re: [mpls-linux-general] about tunneling.

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

On Tue, Jun 26, 2001 at 04:21:43PM +0530, Srinivas Reddy wrote:
> hello all ,
> 
> 	i am getting confussion about tunneling. practically in which case we get 
> tunneling. when we get tunneling we have to send Targetted Hello only(please 
> correct me if i am wrong). Can some body guide me in which cases we will 
> create tunnel especially in the case if we use BGP .

I'm not sure if I understand your question but here is an attempt to
answer it :-)

BTW I hate the term "tunnel" (as in an MPLS tunnel) but since you've used
it, I will use the term and explain what I mean by it.


If we have a signalling protocol set up an LSP to a remote router AND
the signalling protocol associats the remote router ID with the LSP,
then we have an instance where "tunneling" is feasible.  In this case
"tunneling" means that I will treat the remote router as a directly connected
peer (via the LSP) for the purpose of another protocol (signalling OR routing).

Stepping back.  Signalling protocol that setup LSPs and associate router
IDs are:

RSVP-TE (user is responsible for assigning a "destination" address,
in this case the destination address should be the remote router ID)

CR-LDP (the FEC could be the remote router ID /32 or the user is responsible
for assigning a "destination" address, same as for RSVP-TE)

LDP (the FEC is the remote router ID and is normally sent via unsolicited
mode (ala Juniper))

In any of these cases the result will be an LSP which can be associated with
the remote router ID (how the association is made varies).

Once that assocaition is made then the remote router is can be considered
as directly connect (as you would say, a "tunnel" has been created to the
remote router).

Now a second level of protocols can use this "tunnel".  Examples are:

LDP (requires "tunnels" in both directions and then uses targeted hellos
to establish an adj.  The resulting session needs to be associated 
with a "tunnel" for the label mappings to be used)

RSVP-TE (the "tunnel" is treated as one hop in an explicit hop list)

BGP (requires "tunnels" in both directions and each side must advertise their
router ID as the next hop for each NRLI thus BGP can resolve the nexthop via
the "tunnel")

One note about the second level signalling protocols:

The resulting label allocation made by second level signalling protocols are
only known by the end point of the LSP.  Thus packet traversing the LSPs
created by second level signalling protocols have 2 labels on them.  The
top label is for the "tunnel" LSP and the bottom label is for the
LSP created by the second level signalling protocol.

I hope this answers your question :-)

Jim

PS In the above example you can replace router ID with "network wide unique
address that is 'known' to be assocaited with the remote router and is still
valid even when re-routing occurs due to link failures".  :-)

> 
> thanks in advance 
> regards
> lsr
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

Re: [mpls-linux-general] ldp-portable updates

[Venisa C. <vca...@ho...>]

That's it ! It works now :-D Thanks Jim ...

regards,
~Venisa Cabrilla


>From: "James R. Leu" <jl...@mi...>
>Reply-To: jl...@mi...
>To: Venisa Cabrilla <vca...@ho...>
>CC: mpl...@li...
>Subject: Re: [mpls-linux-general] ldp-portable updates
>Date: Tue, 26 Jun 2001 09:57:31 -0500
>
>On Tue, Jun 26, 2001 at 02:28:11PM -0000, Venisa Cabrilla wrote:
> > Hi,
> >
> > Sorry for the late response, I have to go downstair and upstair just to 
>copy
> > the contents of /proc/net/mpls_*, (my lab and my office are not in the 
>same
> > room) but I have done as fast as I can ... :-)
>
>Late response!  I'm used to latency on the order of days (or weeks ;-)
>Your latency is spetacular! :-)
>
> > First, I have to inform you that since I use my computers as routers, I 
>do
> > certainly have at least three ethernet cards in my routers (that is, the
> > topology that I gave you is far from complete, which will explain the
> > important contents of file mpls_in) but I used only the commands that I 
>have
> > described in my last email.
>
>This is cool!  ldp-portable is actually working for someone beside me!
>
> > In router1:
> > mpls_in:
> > 40004000 gen 16 0 POP DLV
> > 40004400 gen 17 0 POP DLV
> > 40004800 gen 18 0 POP DLV
> > 40004c00 gen 19 0 POP DLV
> > 40005000 gen 20 0 POP DLV
> > 40005400 gen 21 0 POP DLV
> > 40005800 gen 22 0 POP DLV
> > mpls_out:
> > 40004003 PUSH(gen 16) SET (eth1)
> > mpls_fec:
> > 40004003 10.0.0.6/32
>
>Ahhhh.  Here is the culprite.
>
> > mpls_labelspace:
> > lo 0
>
>ldp_linux doesn't set the label space for you.  You currently need to 
>execute
>the command 'mplsadm -L eth1:0' for each of the LDP speaking interfaces 
>before
>starting up ldp_linux.
>
>I have a patch from someone at Ericsson that fixes this.
>
>Other then the labelspace problem the rest look just fine :-)
>
>Let me know if this fixes it.
>
>BTW clues to this being the problem would have shown up in the MPLS kernel
>debugging when you tried to ping.
>
>On Router1:
>mplsadm -d
>ping 10.0.0.6
>dmesg
>
>You should see that the packet are being sent via mpls_output()
>
>On Router2:
>ping 10.0.0.5
>
>On Router1:
>dmesg
>
>You should see info about packets arriving on interfaces without a 
>labelspace
>
>Jim
>
> > In router2:
> > mpls_in:
> > 40004000 gen 16 0 POP DLV
> > 40004400 gen 17 0 POP DLV
> > 40004800 gen 18 0 POP DLV
> > 40004c00 gen 19 0 POP DLV
> > 40005000 gen 20 0 POP DLV
> > 40005400 gen 21 0 POP DLV
> > 40005800 gen 22 0 POP DLV
> > mpls_out:
> > 40004004 PUSH(gen 16) SET (eth2)
> > mpls_fec:
> > 40004004 10.0.0.5/32
> > mpls_labelspace:
> > lo 0
> >
> > There is nothing in mpls_tunnel, which is normal.
> >
> >
> > regards,
> > ~Venisa Cabrilla
> >
> >
> >
> > >From: "James R. Leu" <jl...@mi...>
> > >Reply-To: jl...@mi...
> > >To: Venisa Cabrilla <vca...@ho...>
> > >CC: mpl...@li...
> > >Subject: Re: [mpls-linux-general] ldp-portable updates
> > >Date: Tue, 26 Jun 2001 08:45:55 -0500
> > >
> > >On Tue, Jun 26, 2001 at 01:31:11PM -0000, Venisa Cabrilla wrote:
> > > > Hi James,
> > > >
> > > > Thanks for your ldp-portable updates, now I can see that mpls_in,
> > >mpls_out,
> > > > and mpls_fec are not empty.
> > >
> > >Hello,
> > >
> > >I think I know the problem.  It has to do with the fact that LDP is 
>setting
> > >up
> > >LSPs for too many entries in the route table.  Send me the contents of
> > >/proc/net/mpls_* and we can verify this.
> > >
> > >My default policy of "send labels for everything" may not be valid in 
>all
> > >cases.  This will be simple enough to change I just need to figure out 
>what
> > >the
> > >best default policy should be :-(
> > >
> > >I'll try to duplicate this setup here and see if I can figure it out as
> > >well.
> > >
> > >Jim
> > >
> > > >
> > > > I use the simple configuration in file README.sample :
> > > >
> > > >      lo 10.0.0.6                lo 10.0.0.5
> > > >       ---------                  ---------
> > > >      | Router2 |----------------| Router1 |
> > > >       ---------                  ---------
> > > >           |eth2  192.168.12.0/24     |eth1
> > > >      ---------------------------------------
> > > >
> > > > In router1:
> > > > bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > > > bash# route add 10.0.0.6 gw 192.168.12.2
> > > >
> > > > >From inside ldp_linux:
> > > > prompt> add global 10.0.0.5
> > > > prompt> add interface eth1
> > > >
> > > > In router2:
> > > > bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > > > bash# route add 10.0.0.5 gw 192.168.12.1
> > > >
> > > > >From inside ldp_linux:
> > > > prompt> add global 10.0.0.6
> > > > prompt> add interface eth2
> > > >
> > > > After that I can see that mpls_in, mpls_out, and mpls_fec in each 
>router
> > >are
> > > > not empty.
> > > >
> > > > Unfortunately, I cannot ping anymore router1 from router2 and 
>router2
> > >from
> > > > router1. Before I enter ldp_linux, I can ping 10.0.0.5 from router1 
>and
> > >ping
> > > > 10.0.0.6 from router2. However, I can see the MPLS encapsulation 
>using
> > > > ethereal for ping request messages but there is no ping reply 
>message.
> > > >
> > > > Would you mind verify my configuration above and tell me what might 
>be
> > >my
> > > > mistakes ?
> > > >
> > > >
> > > > PS: I have to explicitly add the routes using route commands this 
>time,
> > > > because if I use zebra, I cannot use ethereal (ethereal hangs).
> > > > I still do not know the reason, it might be due to TCP/UDP 
>conflicts.
> > > >
> > > >
> > > > regards,
> > > > ~Venisa Cabrilla
> > > >
> > > >
> > > >
> > > > >From: "James R. Leu" <jl...@mi...>
> > > > >Reply-To: jl...@mi...
> > > > >To: mpl...@li...
> > > > >Subject: [mpls-linux-general] ldp-portable updates
> > > > >Date: Sun, 24 Jun 2001 20:23:13 -0500
> > > > >
> > > > >Sorry to those who have been trying to use ldp-portable from CVS.
> > > > >I finnally have gotten around to fixing and testing ldp-portable.
> > > > >It's almost back to where it was before I started mucking around 
>with
> > >PDU
> > > > >processing.  (interoping with cisco pointed out the need to handle
> > >multiple
> > > > >PDUs per TCP segment and handle one PDU spread across multiple TCP
> > > > >segments)
> > > > >
> > > > >If you tried ldp-portable in the last 3 weeks and it was broken, 
>try
> > > > >updating your CVS client and giving it another go.
> > > > >
> > > > >Sorry for the long delay.
> > > > >Jim
> > > > >--
> > > > >James R. Leu
> > > > >
> > > > >_______________________________________________
> > > > >mpls-linux-general mailing list
> > > > >mpl...@li...
> > > > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > >
> > > >
> > 
> >_________________________________________________________________________
> > > > Get Your Private, Free E-mail from MSN Hotmail at
> > >http://www.hotmail.com.
> > > >
> > > >
> > > > _______________________________________________
> > > > mpls-linux-general mailing list
> > > > mpl...@li...
> > > > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > >
> > >--
> > >James R. Leu
> > >
> > >_______________________________________________
> > >mpls-linux-general mailing list
> > >mpl...@li...
> > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> >
> > 
>_________________________________________________________________________
> > Get Your Private, Free E-mail from MSN Hotmail at 
>http://www.hotmail.com.
>
>--
>James R. Leu

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Re: [mpls-linux-general] ldp-portable updates

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

On Tue, Jun 26, 2001 at 02:28:11PM -0000, Venisa Cabrilla wrote:
> Hi,
> 
> Sorry for the late response, I have to go downstair and upstair just to copy 
> the contents of /proc/net/mpls_*, (my lab and my office are not in the same 
> room) but I have done as fast as I can ... :-)

Late response!  I'm used to latency on the order of days (or weeks ;-)
Your latency is spetacular! :-)

> First, I have to inform you that since I use my computers as routers, I do 
> certainly have at least three ethernet cards in my routers (that is, the 
> topology that I gave you is far from complete, which will explain the 
> important contents of file mpls_in) but I used only the commands that I have 
> described in my last email.

This is cool!  ldp-portable is actually working for someone beside me!

> In router1:
> mpls_in:
> 40004000 gen 16 0 POP DLV
> 40004400 gen 17 0 POP DLV
> 40004800 gen 18 0 POP DLV
> 40004c00 gen 19 0 POP DLV
> 40005000 gen 20 0 POP DLV
> 40005400 gen 21 0 POP DLV
> 40005800 gen 22 0 POP DLV
> mpls_out:
> 40004003 PUSH(gen 16) SET (eth1)
> mpls_fec:
> 40004003 10.0.0.6/32

Ahhhh.  Here is the culprite.

> mpls_labelspace:
> lo 0

ldp_linux doesn't set the label space for you.  You currently need to execute
the command 'mplsadm -L eth1:0' for each of the LDP speaking interfaces before
starting up ldp_linux.

I have a patch from someone at Ericsson that fixes this.

Other then the labelspace problem the rest look just fine :-)

Let me know if this fixes it.

BTW clues to this being the problem would have shown up in the MPLS kernel
debugging when you tried to ping.

On Router1:
mplsadm -d
ping 10.0.0.6
dmesg

You should see that the packet are being sent via mpls_output()

On Router2:
ping 10.0.0.5

On Router1:
dmesg

You should see info about packets arriving on interfaces without a labelspace

Jim

> In router2:
> mpls_in:
> 40004000 gen 16 0 POP DLV
> 40004400 gen 17 0 POP DLV
> 40004800 gen 18 0 POP DLV
> 40004c00 gen 19 0 POP DLV
> 40005000 gen 20 0 POP DLV
> 40005400 gen 21 0 POP DLV
> 40005800 gen 22 0 POP DLV
> mpls_out:
> 40004004 PUSH(gen 16) SET (eth2)
> mpls_fec:
> 40004004 10.0.0.5/32
> mpls_labelspace:
> lo 0
> 
> There is nothing in mpls_tunnel, which is normal.
> 
> 
> regards,
> ~Venisa Cabrilla
> 
> 
> 
> >From: "James R. Leu" <jl...@mi...>
> >Reply-To: jl...@mi...
> >To: Venisa Cabrilla <vca...@ho...>
> >CC: mpl...@li...
> >Subject: Re: [mpls-linux-general] ldp-portable updates
> >Date: Tue, 26 Jun 2001 08:45:55 -0500
> >
> >On Tue, Jun 26, 2001 at 01:31:11PM -0000, Venisa Cabrilla wrote:
> > > Hi James,
> > >
> > > Thanks for your ldp-portable updates, now I can see that mpls_in, 
> >mpls_out,
> > > and mpls_fec are not empty.
> >
> >Hello,
> >
> >I think I know the problem.  It has to do with the fact that LDP is setting 
> >up
> >LSPs for too many entries in the route table.  Send me the contents of
> >/proc/net/mpls_* and we can verify this.
> >
> >My default policy of "send labels for everything" may not be valid in all
> >cases.  This will be simple enough to change I just need to figure out what 
> >the
> >best default policy should be :-(
> >
> >I'll try to duplicate this setup here and see if I can figure it out as 
> >well.
> >
> >Jim
> >
> > >
> > > I use the simple configuration in file README.sample :
> > >
> > >      lo 10.0.0.6                lo 10.0.0.5
> > >       ---------                  ---------
> > >      | Router2 |----------------| Router1 |
> > >       ---------                  ---------
> > >           |eth2  192.168.12.0/24     |eth1
> > >      ---------------------------------------
> > >
> > > In router1:
> > > bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > > bash# route add 10.0.0.6 gw 192.168.12.2
> > >
> > > >From inside ldp_linux:
> > > prompt> add global 10.0.0.5
> > > prompt> add interface eth1
> > >
> > > In router2:
> > > bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > > bash# route add 10.0.0.5 gw 192.168.12.1
> > >
> > > >From inside ldp_linux:
> > > prompt> add global 10.0.0.6
> > > prompt> add interface eth2
> > >
> > > After that I can see that mpls_in, mpls_out, and mpls_fec in each router 
> >are
> > > not empty.
> > >
> > > Unfortunately, I cannot ping anymore router1 from router2 and router2 
> >from
> > > router1. Before I enter ldp_linux, I can ping 10.0.0.5 from router1 and 
> >ping
> > > 10.0.0.6 from router2. However, I can see the MPLS encapsulation using
> > > ethereal for ping request messages but there is no ping reply message.
> > >
> > > Would you mind verify my configuration above and tell me what might be 
> >my
> > > mistakes ?
> > >
> > >
> > > PS: I have to explicitly add the routes using route commands this time,
> > > because if I use zebra, I cannot use ethereal (ethereal hangs).
> > > I still do not know the reason, it might be due to TCP/UDP conflicts.
> > >
> > >
> > > regards,
> > > ~Venisa Cabrilla
> > >
> > >
> > >
> > > >From: "James R. Leu" <jl...@mi...>
> > > >Reply-To: jl...@mi...
> > > >To: mpl...@li...
> > > >Subject: [mpls-linux-general] ldp-portable updates
> > > >Date: Sun, 24 Jun 2001 20:23:13 -0500
> > > >
> > > >Sorry to those who have been trying to use ldp-portable from CVS.
> > > >I finnally have gotten around to fixing and testing ldp-portable.
> > > >It's almost back to where it was before I started mucking around with 
> >PDU
> > > >processing.  (interoping with cisco pointed out the need to handle 
> >multiple
> > > >PDUs per TCP segment and handle one PDU spread across multiple TCP
> > > >segments)
> > > >
> > > >If you tried ldp-portable in the last 3 weeks and it was broken, try
> > > >updating your CVS client and giving it another go.
> > > >
> > > >Sorry for the long delay.
> > > >Jim
> > > >--
> > > >James R. Leu
> > > >
> > > >_______________________________________________
> > > >mpls-linux-general mailing list
> > > >mpl...@li...
> > > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > >
> > > 
> >_________________________________________________________________________
> > > Get Your Private, Free E-mail from MSN Hotmail at 
> >http://www.hotmail.com.
> > >
> > >
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> >
> >--
> >James R. Leu
> >
> >_______________________________________________
> >mpls-linux-general mailing list
> >mpl...@li...
> >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 
> _________________________________________________________________________
> Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

-- 
James R. Leu

[mpls-linux-general] mpls-tc patch

[Stefano A. <avi...@li...>]

Hi Steven and all,

I would greatly appreciate if someone could say me when the new release =
of the mpls-tc patch will be available
( the implementation of Diffserv on a mpls-linux network is the subject =
of my thesis )

Thanks a lot,
Stefano

Re: [mpls-linux-general] ldp-portable updates

[Venisa C. <vca...@ho...>]

Hi,

Sorry for the late response, I have to go downstair and upstair just to copy 
the contents of /proc/net/mpls_*, (my lab and my office are not in the same 
room) but I have done as fast as I can ... :-)

First, I have to inform you that since I use my computers as routers, I do 
certainly have at least three ethernet cards in my routers (that is, the 
topology that I gave you is far from complete, which will explain the 
important contents of file mpls_in) but I used only the commands that I have 
described in my last email.

In router1:
mpls_in:
40004000 gen 16 0 POP DLV
40004400 gen 17 0 POP DLV
40004800 gen 18 0 POP DLV
40004c00 gen 19 0 POP DLV
40005000 gen 20 0 POP DLV
40005400 gen 21 0 POP DLV
40005800 gen 22 0 POP DLV
mpls_out:
40004003 PUSH(gen 16) SET (eth1)
mpls_fec:
40004003 10.0.0.6/32
mpls_labelspace:
lo 0

In router2:
mpls_in:
40004000 gen 16 0 POP DLV
40004400 gen 17 0 POP DLV
40004800 gen 18 0 POP DLV
40004c00 gen 19 0 POP DLV
40005000 gen 20 0 POP DLV
40005400 gen 21 0 POP DLV
40005800 gen 22 0 POP DLV
mpls_out:
40004004 PUSH(gen 16) SET (eth2)
mpls_fec:
40004004 10.0.0.5/32
mpls_labelspace:
lo 0

There is nothing in mpls_tunnel, which is normal.


regards,
~Venisa Cabrilla



>From: "James R. Leu" <jl...@mi...>
>Reply-To: jl...@mi...
>To: Venisa Cabrilla <vca...@ho...>
>CC: mpl...@li...
>Subject: Re: [mpls-linux-general] ldp-portable updates
>Date: Tue, 26 Jun 2001 08:45:55 -0500
>
>On Tue, Jun 26, 2001 at 01:31:11PM -0000, Venisa Cabrilla wrote:
> > Hi James,
> >
> > Thanks for your ldp-portable updates, now I can see that mpls_in, 
>mpls_out,
> > and mpls_fec are not empty.
>
>Hello,
>
>I think I know the problem.  It has to do with the fact that LDP is setting 
>up
>LSPs for too many entries in the route table.  Send me the contents of
>/proc/net/mpls_* and we can verify this.
>
>My default policy of "send labels for everything" may not be valid in all
>cases.  This will be simple enough to change I just need to figure out what 
>the
>best default policy should be :-(
>
>I'll try to duplicate this setup here and see if I can figure it out as 
>well.
>
>Jim
>
> >
> > I use the simple configuration in file README.sample :
> >
> >      lo 10.0.0.6                lo 10.0.0.5
> >       ---------                  ---------
> >      | Router2 |----------------| Router1 |
> >       ---------                  ---------
> >           |eth2  192.168.12.0/24     |eth1
> >      ---------------------------------------
> >
> > In router1:
> > bash# ifconfig lo 10.0.0.5 netmask 255.255.255.255
> > bash# route add 10.0.0.6 gw 192.168.12.2
> >
> > >From inside ldp_linux:
> > prompt> add global 10.0.0.5
> > prompt> add interface eth1
> >
> > In router2:
> > bash# ifconfig lo 10.0.0.6 netmask 255.255.255.255
> > bash# route add 10.0.0.5 gw 192.168.12.1
> >
> > >From inside ldp_linux:
> > prompt> add global 10.0.0.6
> > prompt> add interface eth2
> >
> > After that I can see that mpls_in, mpls_out, and mpls_fec in each router 
>are
> > not empty.
> >
> > Unfortunately, I cannot ping anymore router1 from router2 and router2 
>from
> > router1. Before I enter ldp_linux, I can ping 10.0.0.5 from router1 and 
>ping
> > 10.0.0.6 from router2. However, I can see the MPLS encapsulation using
> > ethereal for ping request messages but there is no ping reply message.
> >
> > Would you mind verify my configuration above and tell me what might be 
>my
> > mistakes ?
> >
> >
> > PS: I have to explicitly add the routes using route commands this time,
> > because if I use zebra, I cannot use ethereal (ethereal hangs).
> > I still do not know the reason, it might be due to TCP/UDP conflicts.
> >
> >
> > regards,
> > ~Venisa Cabrilla
> >
> >
> >
> > >From: "James R. Leu" <jl...@mi...>
> > >Reply-To: jl...@mi...
> > >To: mpl...@li...
> > >Subject: [mpls-linux-general] ldp-portable updates
> > >Date: Sun, 24 Jun 2001 20:23:13 -0500
> > >
> > >Sorry to those who have been trying to use ldp-portable from CVS.
> > >I finnally have gotten around to fixing and testing ldp-portable.
> > >It's almost back to where it was before I started mucking around with 
>PDU
> > >processing.  (interoping with cisco pointed out the need to handle 
>multiple
> > >PDUs per TCP segment and handle one PDU spread across multiple TCP
> > >segments)
> > >
> > >If you tried ldp-portable in the last 3 weeks and it was broken, try
> > >updating your CVS client and giving it another go.
> > >
> > >Sorry for the long delay.
> > >Jim
> > >--
> > >James R. Leu
> > >
> > >_______________________________________________
> > >mpls-linux-general mailing list
> > >mpl...@li...
> > >http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> >
> > 
>_________________________________________________________________________
> > Get Your Private, Free E-mail from MSN Hotmail at 
>http://www.hotmail.com.
> >
> >
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > http://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
>--
>James R. Leu
>
>_______________________________________________
>mpls-linux-general mailing list
>mpl...@li...
>http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.