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

Re: [mpls-linux-general] Follow on test of 1.956 - 1.955 patches as applied to kernel and iproute

[Chris R. <Chr...@nr...>]

More test results:

Kernel:  2.6.20-1.2316fc5.mpls.1.956 built by James (or Marco aka 
binaries pulled from mirror)
Iproute2:  1.956 iproute2-2.6.16-060323, I built)

Still get the same failure as discussed previously......  this confirms 
the kernel patches for anything above 
kernel-2.6.18-1.2257.fc5.mpls.1.955 distro isnt working correctly

...Chris
> James, Marco
>
> This last test used the kernel binaries for 1.955 and iproute2 1.956 
> binaries I built.  Now the iproute commands work, see below, ergo 
> something isnt right with the kernel patches.  Any ideas.....
>
> # ./mpls/mpls nhlfe add key 0 instructions push gen 16 nexthop eth0 ipv4 
> 10.128.142.1
> NHLFE entry key 0x00000002 mtu 0 propagate_ttl
>          (0 bytes, 0 pkts)
> # ./ip/ip route add 192.168.120.0/24 via 10.128.142.1 mpls 0x2
> # ./ip/ip route show
> 10.10.1.25 dev tun10  proto kernel  scope link  src 10.10.1.26
> 10.3.1.9 dev tun3  proto kernel  scope link  src 10.3.1.10
> 10.5.1.9 dev tun5  proto kernel  scope link  src 10.5.1.10
> 10.8.1.0/24 via 10.3.1.9 dev tun3
> 10.10.1.0/24 via 10.10.1.25 dev tun10
> 192.168.120.0/24 via 10.128.142.1 dev eth0 mpls 0x2
> 10.5.1.0/24 via 10.5.1.9 dev tun5
> 10.3.1.0/24 via 10.3.1.9 dev tun3
> 192.168.122.0/24 dev virbr0  proto kernel  scope link  src 192.168.122.1
> 10.128.142.0/24 dev eth0  proto kernel  scope link  src 10.128.142.196
> default via 10.128.142.1 dev eth0
>
>
> Chris Robson wrote:
>   
>> James, Marco
>>
>> I installed binaries (kernel) 2.6.20-1.2316fc5.mpls.1.956 and binaries 
>> (iproute) iproute2-2.6.16-060323 with the MPLS patches and get the 
>> exact same error as before with the kernel 2.6.21 builds.    The only 
>> thing I can think is because I'm running "ip" and "mpls" commands from 
>> the build directories and not the install directories is causing the 
>> problem?  What do you guys think?
>>
>> ...Chris
>>
>> Chris Robson wrote:
>>     
>>> Gentlemen
>>>
>>> I was able to build the 2.6.21 kernel with the 1.956 patches but had 
>>> to make some small changes enclosed below. Further I was able to 
>>> issue the example commands however as you can see the "ip route show" 
>>> command dumps a garbage string:
>>>
>>> # ./mpls/mpls nhlfe add key 0 instructions push gen 16 nexthop eth0 
>>> ipv4 10.128.142.1
>>> NHLFE entry key 0x00000002 mtu 0 propagate_ttl
>>> (0 bytes, 0 pkts)
>>> # ./ip/ip route add 192.168.120.0/24 via 10.128.142.1 mpls 0x2
>>>
>>> # ./ip/ip route show
>>> 10.10.1.25 dev tun10 shim � 0x080004fffd01000 ................ 
>>> 000000000 proto kernel scope link src 10.10.1.26
>>> :
>>> :
>>> 10.3.1.9 dev tun3 shim � 0x08000400070000003 ................. 
>>> 00000000000 proto kernel scope link src 10.5.1.10
>>> 10.128.142.0/24 dev eth0 shim � 0x proto kernel scope link src 
>>> 10.128.142.196
>>> default via 10.128.142.1 dev eth0 shim � 0x
>>>
>>> The following are the changes I had to make to the linux-kernel.diif 
>>> patches to get the kernel to compile and run. As for the iproute2 
>>> files, I just used James's iproute2 sources.
>>>
>>> --- /develop/MPLS/mpls-linux-1.956/patches/linux-kernel.diff 
>>> 2007-05-04 00:51:06.000000000 -0400
>>> +++ MPLS-MULTIPLE-PATCHES.linux-kernel.diff.CHRIS-MODS.28Jun07.patch 
>>> 2007-06-28 12:54:08.000000000 -0400
>>> @@ -311,10 +311,10 @@
>>> diff -uNr --exclude=.p4config linux-kernel/include/linux/netdevice.h 
>>> mpls-kernel/include/linux/netdevice.h
>>> --- linux-kernel/include/linux/netdevice.h 2007-03-24 
>>> 22:55:23.000000000 -0500
>>> +++ mpls-kernel/include/linux/netdevice.h 2007-03-24 
>>> 23:28:29.000000000 -0500
>>> -@@ -398,6 +398,7 @@
>>> - void *ip6_ptr; /* IPv6 specific data */
>>> - void *ec_ptr; /* Econet specific data */
>>> +@@ -402,6 +402,7 @@
>>> void *ax25_ptr; /* AX.25 specific data */
>>> + struct wireless_dev *ieee80211_ptr; /* IEEE 802.11 specific data,
>>> + assign before registering */
>>> + void *mpls_ptr; /* MPLS specific data */
>>>
>>> /*
>>> @@ -2638,12 +2638,12 @@
>>> +
>>> +static int __init init(void)
>>> +{
>>> -+ return ipt_register_target(&ipt_mpls_reg);
>>> ++ return xt_register_target(&ipt_mpls_reg);
>>> +}
>>> +
>>> +static void __exit fini(void)
>>> +{
>>> -+ ipt_unregister_target(&ipt_mpls_reg);
>>> ++ xt_unregister_target(&ipt_mpls_reg);
>>> +}
>>> +
>>> +module_init(init);
>>> @@ -2737,9 +2737,9 @@
>>>
>>> +EXPORT_SYMBOL(ipv6_addr_type);
>>> +EXPORT_SYMBOL(ipv6_rcv);
>>> - EXPORT_SYMBOL(__ipv6_addr_type);
>>> EXPORT_SYMBOL(icmpv6_send);
>>> EXPORT_SYMBOL(icmpv6_statistics);
>>> + EXPORT_SYMBOL(icmpv6_err_convert);
>>> diff -uNr --exclude=.p4config linux-kernel/net/ipv6/Kconfig 
>>> mpls-kernel/net/ipv6/Kconfig
>>> --- linux-kernel/net/ipv6/Kconfig 2007-03-24 22:55:24.000000000 -0500
>>> +++ mpls-kernel/net/ipv6/Kconfig 2007-03-24 23:29:41.000000000 -0500
>>> @@ -3033,12 +3033,12 @@
>>> +
>>> +static int __init init(void)
>>> +{
>>> -+ return ip6t_register_target(&ip6t_mpls_reg);
>>> ++ return xt_register_target(&ip6t_mpls_reg);
>>> +}
>>> +
>>> +static void __exit fini(void)
>>> +{
>>> -+ ip6t_unregister_target(&ip6t_mpls_reg);
>>> ++ xt_unregister_target(&ip6t_mpls_reg);
>>> +}
>>> +
>>> +module_init(init);
>>> @@ -10063,7 +10063,7 @@
>>> diff -uNr --exclude=.p4config linux-kernel/net/mpls/mpls_tunnel.c 
>>> mpls-kernel/net/mpls/mpls_tunnel.c
>>> --- linux-kernel/net/mpls/mpls_tunnel.c 1969-12-31 18:00:00.000000000 
>>> -0600
>>> +++ mpls-kernel/net/mpls/mpls_tunnel.c 2007-02-28 22:07:06.000000000 
>>> -0600
>>> -@@ -0,0 +1,531 @@
>>> +@@ -0,0 +1,532 @@
>>> +/***************************************************************************** 
>>>
>>> + * MPLS
>>> + * An implementation of the MPLS (MultiProtocol Label
>>> @@ -10457,7 +10457,8 @@
>>> + break;
>>> + dev = mpls_mtp2dev(mtp);
>>> + }
>>> -+ retval = unregister_netdevice(dev);
>>> ++ unregister_netdevice(dev);
>>> ++ retval = 0;
>>> + break;
>>> + default:
>>> + retval = -EINVAL;
>>>
>>>
>>>
>>>       
>>     
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
>
>
>   

[mpls-linux-general] Follow on test of 1.956 - 1.955 patches as applied to kernel and iproute

[Chris R. <Chr...@nr...>]

James, Marco

This last test used the kernel binaries for 1.955 and iproute2 1.956 
binaries I built.  Now the iproute commands work, see below, ergo 
something isnt right with the kernel patches.  Any ideas.....

# ./mpls/mpls nhlfe add key 0 instructions push gen 16 nexthop eth0 ipv4 
10.128.142.1
NHLFE entry key 0x00000002 mtu 0 propagate_ttl
         (0 bytes, 0 pkts)
# ./ip/ip route add 192.168.120.0/24 via 10.128.142.1 mpls 0x2
# ./ip/ip route show
10.10.1.25 dev tun10  proto kernel  scope link  src 10.10.1.26
10.3.1.9 dev tun3  proto kernel  scope link  src 10.3.1.10
10.5.1.9 dev tun5  proto kernel  scope link  src 10.5.1.10
10.8.1.0/24 via 10.3.1.9 dev tun3
10.10.1.0/24 via 10.10.1.25 dev tun10
192.168.120.0/24 via 10.128.142.1 dev eth0 mpls 0x2
10.5.1.0/24 via 10.5.1.9 dev tun5
10.3.1.0/24 via 10.3.1.9 dev tun3
192.168.122.0/24 dev virbr0  proto kernel  scope link  src 192.168.122.1
10.128.142.0/24 dev eth0  proto kernel  scope link  src 10.128.142.196
default via 10.128.142.1 dev eth0


Chris Robson wrote:
>
> James, Marco
>
> I installed binaries (kernel) 2.6.20-1.2316fc5.mpls.1.956 and binaries 
> (iproute) iproute2-2.6.16-060323 with the MPLS patches and get the 
> exact same error as before with the kernel 2.6.21 builds.    The only 
> thing I can think is because I'm running "ip" and "mpls" commands from 
> the build directories and not the install directories is causing the 
> problem?  What do you guys think?
>
> ...Chris
>
> Chris Robson wrote:
>>
>> Gentlemen
>>
>> I was able to build the 2.6.21 kernel with the 1.956 patches but had 
>> to make some small changes enclosed below. Further I was able to 
>> issue the example commands however as you can see the "ip route show" 
>> command dumps a garbage string:
>>
>> # ./mpls/mpls nhlfe add key 0 instructions push gen 16 nexthop eth0 
>> ipv4 10.128.142.1
>> NHLFE entry key 0x00000002 mtu 0 propagate_ttl
>> (0 bytes, 0 pkts)
>> # ./ip/ip route add 192.168.120.0/24 via 10.128.142.1 mpls 0x2
>>
>> # ./ip/ip route show
>> 10.10.1.25 dev tun10 shim � 0x080004fffd01000 ................ 
>> 000000000 proto kernel scope link src 10.10.1.26
>> :
>> :
>> 10.3.1.9 dev tun3 shim � 0x08000400070000003 ................. 
>> 00000000000 proto kernel scope link src 10.5.1.10
>> 10.128.142.0/24 dev eth0 shim � 0x proto kernel scope link src 
>> 10.128.142.196
>> default via 10.128.142.1 dev eth0 shim � 0x
>>
>> The following are the changes I had to make to the linux-kernel.diif 
>> patches to get the kernel to compile and run. As for the iproute2 
>> files, I just used James's iproute2 sources.
>>
>> --- /develop/MPLS/mpls-linux-1.956/patches/linux-kernel.diff 
>> 2007-05-04 00:51:06.000000000 -0400
>> +++ MPLS-MULTIPLE-PATCHES.linux-kernel.diff.CHRIS-MODS.28Jun07.patch 
>> 2007-06-28 12:54:08.000000000 -0400
>> @@ -311,10 +311,10 @@
>> diff -uNr --exclude=.p4config linux-kernel/include/linux/netdevice.h 
>> mpls-kernel/include/linux/netdevice.h
>> --- linux-kernel/include/linux/netdevice.h 2007-03-24 
>> 22:55:23.000000000 -0500
>> +++ mpls-kernel/include/linux/netdevice.h 2007-03-24 
>> 23:28:29.000000000 -0500
>> -@@ -398,6 +398,7 @@
>> - void *ip6_ptr; /* IPv6 specific data */
>> - void *ec_ptr; /* Econet specific data */
>> +@@ -402,6 +402,7 @@
>> void *ax25_ptr; /* AX.25 specific data */
>> + struct wireless_dev *ieee80211_ptr; /* IEEE 802.11 specific data,
>> + assign before registering */
>> + void *mpls_ptr; /* MPLS specific data */
>>
>> /*
>> @@ -2638,12 +2638,12 @@
>> +
>> +static int __init init(void)
>> +{
>> -+ return ipt_register_target(&ipt_mpls_reg);
>> ++ return xt_register_target(&ipt_mpls_reg);
>> +}
>> +
>> +static void __exit fini(void)
>> +{
>> -+ ipt_unregister_target(&ipt_mpls_reg);
>> ++ xt_unregister_target(&ipt_mpls_reg);
>> +}
>> +
>> +module_init(init);
>> @@ -2737,9 +2737,9 @@
>>
>> +EXPORT_SYMBOL(ipv6_addr_type);
>> +EXPORT_SYMBOL(ipv6_rcv);
>> - EXPORT_SYMBOL(__ipv6_addr_type);
>> EXPORT_SYMBOL(icmpv6_send);
>> EXPORT_SYMBOL(icmpv6_statistics);
>> + EXPORT_SYMBOL(icmpv6_err_convert);
>> diff -uNr --exclude=.p4config linux-kernel/net/ipv6/Kconfig 
>> mpls-kernel/net/ipv6/Kconfig
>> --- linux-kernel/net/ipv6/Kconfig 2007-03-24 22:55:24.000000000 -0500
>> +++ mpls-kernel/net/ipv6/Kconfig 2007-03-24 23:29:41.000000000 -0500
>> @@ -3033,12 +3033,12 @@
>> +
>> +static int __init init(void)
>> +{
>> -+ return ip6t_register_target(&ip6t_mpls_reg);
>> ++ return xt_register_target(&ip6t_mpls_reg);
>> +}
>> +
>> +static void __exit fini(void)
>> +{
>> -+ ip6t_unregister_target(&ip6t_mpls_reg);
>> ++ xt_unregister_target(&ip6t_mpls_reg);
>> +}
>> +
>> +module_init(init);
>> @@ -10063,7 +10063,7 @@
>> diff -uNr --exclude=.p4config linux-kernel/net/mpls/mpls_tunnel.c 
>> mpls-kernel/net/mpls/mpls_tunnel.c
>> --- linux-kernel/net/mpls/mpls_tunnel.c 1969-12-31 18:00:00.000000000 
>> -0600
>> +++ mpls-kernel/net/mpls/mpls_tunnel.c 2007-02-28 22:07:06.000000000 
>> -0600
>> -@@ -0,0 +1,531 @@
>> +@@ -0,0 +1,532 @@
>> +/***************************************************************************** 
>>
>> + * MPLS
>> + * An implementation of the MPLS (MultiProtocol Label
>> @@ -10457,7 +10457,8 @@
>> + break;
>> + dev = mpls_mtp2dev(mtp);
>> + }
>> -+ retval = unregister_netdevice(dev);
>> ++ unregister_netdevice(dev);
>> ++ retval = 0;
>> + break;
>> + default:
>> + retval = -EINVAL;
>>
>>
>>
>
>

[mpls-linux-general] 1.956 testing volunteer

[Chris R. <Chr...@nr...>]

James

Where can I get 1.956 to start testing.  Further, I've built 1.955 for 
F-7 with kernel 2.6.21 but it has major issues.  I suspect the problems 
are in fib_semantics.c and route.c as those are the files requiring the 
most dramatic changes.  And there are issues with iproute2 (latest) as 
well...I'll try collecting more details over the next few days.....

...Chris

[mpls-linux-general] targeted LDP session

[Xu Y. <ya...@gm...>]

Hello,

Could anyone please tell me how to configure a targeted LDP session?
I have tried to search commands in vtysh but I couldn't find one :-(

Thanks!
Yao

Re: [mpls-linux-general] on equal cost multi path

[Xu Y. <ya...@gm...>]

Hello,

I've tried with quagga 0.99.7 without MPLS and I got the same result.
I guess it's due to the inconsistency between the configuration of the
kernel and that of quagga.
Any comments are welcome.

Thanks.


2007/6/25, James R. Leu <jl...@mi...>:
>
> This might be an error with the changes I have made to the FIB/RIB to
> support
> MPLS.
>
> I would suggest trying the same test on a non-mpls version of quagga and
> see if you get the same results.
>
> On Sun, Jun 24, 2007 at 08:06:04PM +0200, Xu Yao wrote:
> > Hello,
> >
> > I have a problem with OSPF equal cost multi path.
> > In my test, I noticed that ospfd had calculated two routes
> > N    172.16.20.0/24        [30] area: 0.0.0.0
> >                                            via 10.0.2.3, eth0
> >                                            via 10.0.1.5, eth3
> > but none of them was written into the FIB by zebra
> > O>  172.16.20.0/24 [110/30] via 10.0.2.3, eth0, 00:00:39
> >                                            via 10.0.1.5, eth3, 00:00:39
> >
> > I am now using
> > quagga-0.99.6-01.fc5.mpls.1.955d.i386.rpm<
> http://downloads.sourceforge.net/mpls-linux/quagga-0.99.6-01.fc5.mpls.1.955d.i386.rpm?modtime=1174359745&big_mirror=0
> >
> > downloaded from the mpls-linux website and I didn't enable ECMP when I
> > compiled the kernel.
> >
> > I am expecting the behaviour that one of the equal cost routes be
> written in
> > the FIB, what should I do?
> >
> > Thanks
> > Yao
>
> >
> -------------------------------------------------------------------------
> > This SF.net email is sponsored by DB2 Express
> > Download DB2 Express C - the FREE version of DB2 express and take
> > control of your XML. No limits. Just data. Click to get it now.
> > http://sourceforge.net/powerbar/db2/
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
>
> --
> James R. Leu
> jl...@mi...
>
>

Re: [mpls-linux-general] on equal cost multi path

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

This might be an error with the changes I have made to the FIB/RIB to suppo=
rt
MPLS.

I would suggest trying the same test on a non-mpls version of quagga and
see if you get the same results.

On Sun, Jun 24, 2007 at 08:06:04PM +0200, Xu Yao wrote:
> Hello,
>=20
> I have a problem with OSPF equal cost multi path.
> In my test, I noticed that ospfd had calculated two routes
> N    172.16.20.0/24        [30] area: 0.0.0.0
>                                            via 10.0.2.3, eth0
>                                            via 10.0.1.5, eth3
> but none of them was written into the FIB by zebra
> O>  172.16.20.0/24 [110/30] via 10.0.2.3, eth0, 00:00:39
>                                            via 10.0.1.5, eth3, 00:00:39
>=20
> I am now using
> quagga-0.99.6-01.fc5.mpls.1.955d.i386.rpm<http://downloads.sourceforge.ne=
t/mpls-linux/quagga-0.99.6-01.fc5.mpls.1.955d.i386.rpm?modtime=3D1174359745=
&big_mirror=3D0>
> downloaded from the mpls-linux website and I didn't enable ECMP when I
> compiled the kernel.
>=20
> I am expecting the behaviour that one of the equal cost routes be written=
 in
> the FIB, what should I do?
>=20
> Thanks
> Yao

> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general


--=20
James R. Leu
jl...@mi...

[mpls-linux-general] on equal cost multi path

[Xu Y. <ya...@gm...>]

Hello,

I have a problem with OSPF equal cost multi path.
In my test, I noticed that ospfd had calculated two routes
N    172.16.20.0/24        [30] area: 0.0.0.0
                                            via 10.0.2.3, eth0
                                            via 10.0.1.5, eth3
but none of them was written into the FIB by zebra
O>  172.16.20.0/24 [110/30] via 10.0.2.3, eth0, 00:00:39
                                            via 10.0.1.5, eth3, 00:00:39

I am now using
quagga-0.99.6-01.fc5.mpls.1.955d.i386.rpm<http://downloads.sourceforge.net/mpls-linux/quagga-0.99.6-01.fc5.mpls.1.955d.i386.rpm?modtime=1174359745&big_mirror=0>
downloaded from the mpls-linux website and I didn't enable ECMP when I
compiled the kernel.

I am expecting the behaviour that one of the equal cost routes be written in
the FIB, what should I do?

Thanks
Yao

Re: [mpls-linux-general] How ldp assigns labes

[Dragoescu A. <dra...@gm...>]

On 6/23/07, Marco Schmidt <mar...@s2...> wrote:
>
> You mean a setup like this?
>
>   ---B---
>       /         \
>     A< >D
>       \         /
>   ---C---


ok with this setup , ospf will distribute 2 routes with destination router D
and will mark one of them as the best.If i send mpls trafic to router D ,
that traffic will follow that route let's say trough router B.My question
is ,  if disconnect one link of router B , after a while traffic will be
sent trough router C?
10x

-- 
Alexandru Dragoescu
dra...@gm...
QA engineer
ITCNETWORKS Romania

Re: [mpls-linux-general] How ldp assigns labes

[Dragoescu A. <dra...@gm...>]

does ldp implementation support loop detection?

On 6/23/07, Marco Schmidt <mar...@s2...> wrote:
>
> You mean a setup like this?
>
>   ---B---
>       /         \
>     A< >D
>       \         /
>   ---C---
>
> This scenario doesn't work since you change the link cost (in OSPF) on one
> link. The LDP-implementation doesn't support equal cost multipaths
> (ECMPs).
>
> Routes are redistributed by the zebra routing manager, you may look at the
> resources of quagga.net to see how the protocol deamons interact with
> each other.
>
> >
> >
> >
> > Hy,
> >
> >
> >
> > I have a question ? How ldp assigns labels.Have this setup:
> >
> >
> >
> > B    AD
> >
> >
> >
> >
> >
> > A connected with B and C and D also
> >
> > Ok,after ospf will distribute routes , how ldo assigns labels.
> >
> > On LSR A Ospf will distribute 2 routes with destination D.
> >
> > LDP will create 2 LSP with destination D ?
> >
> >
> >
> > As implementation how LDP will assign labes? It looks on ospf
> > routing table and for any destination will assign a new label?
> >
> >
> >
> > 10x,
> >
> > AlexD
> >
> >
> >
>
> --
> Marco Schmidt
> mailto:mar...@s2...
>
>
>


-- 
Alexandru Dragoescu
dra...@gm...
QA engineer
ITCNETWORKS Romania

Re: [mpls-linux-general] How ldp assigns labes

[Marco S. <mar...@s2...>]

You mean a setup like this?

=A0=A0=A0=A0=A0 =A0 ---B---
      /         \
    A<=A0=A0=A0=A0=A0 =A0=A0=A0=A0=A0>D
      \         /
 =A0=A0=A0=A0=A0=A0 ---C---

This scenario doesn't work since you change the link cost (in OSPF) on one
link. The LDP-implementation doesn't support equal cost multipaths (ECMPs).

Routes are redistributed by the zebra routing manager, you may look at the
resources of quagga.net to see how the protocol deamons interact with
each other.

>    =20
>  =20
>  =20
> Hy,
>  =20
> =A0
>  =20
> I have a question ? How ldp assigns labels.Have this setup:
>  =20
> =A0
>  =20
> =A0=A0=A0=A0=A0=A0=A0B    A=A0=A0=A0=A0D
>
> =A0=A0=A0=A0=A0=A0=A0
>  =20
> =A0
>  =20
> A connected with B and C and D also=20
>  =20
> Ok,after ospf will distribute routes , how ldo assigns labels.
>  =20
> On LSR A Ospf will distribute 2 routes with destination D.
>  =20
> LDP will create 2 LSP with destination D ?
>  =20
> =A0
>  =20
> As implementation how LDP will assign labes? It looks on ospf
> routing table and for any destination will assign a new label?
>  =20
> =A0
>  =20
> 10x,
>  =20
> AlexD
>  =20
>  =20
>    =20

--=20
Marco Schmidt
mailto:mar...@s2...

[mpls-linux-general] How ldp assigns labes

[Dragoescu A. <dra...@gm...>]

Hy,

 

I have a question ? How ldp assigns labels.Have this setup:

 

            B

A                      D

            C

 

A connected with B and C and D also 

Ok,after ospf will distribute routes , how ldo assigns labels.

On LSR A Ospf will distribute 2 routes with destination D.

LDP will create 2 LSP with destination D ?

 

As implementation how LDP will assign labes? It looks on ospf routing table
and for any destination will assign a new label?

 

10x,

AlexD

Re: [mpls-linux-general] vtysh not work

[Xu Y. <ya...@gm...>]

Hello,

I have a simple workaround.
Create a symbolic link libtermcap.so.2 -> libncursesw.so.5.4
I don't know if there's any bad effect but at least it works.

Good luck

2007/6/23, elc...@en... <elc...@en...>:
>
> Hello,
>
> I installed the quagga package from sourceforge and then I typed the
> command "vtysh" but this isn't work properly and I recieved this error:
> vtysh: error while loading shared libraries: libtermcap.so.2: cannot open
> shared object file : No such file or directory
>
> for information:
> I use the quagga.99.6 version
> and the netkit uml kernel 2.6.18.8 debian which I compiled after I patched
> it with mpls-linux-1.955 patch form sourceforge
>
> please help me to continue this work
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>

Re: [mpls-linux-general] vtysh not work

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

I know this is an obvious question, but do you have libtermcap
install on your system?

On Sat, Jun 23, 2007 at 12:17:39AM +0200, elc...@en... wrote:
> Hello,
>=20
> I installed the quagga package from sourceforge and then I typed the
> command "vtysh" but this isn't work properly and I recieved this error:
> vtysh: error while loading shared libraries: libtermcap.so.2: cannot open
> shared object file : No such file or directory
>=20
> for information:
> I use the quagga.99.6 version
> and the netkit uml kernel 2.6.18.8 debian which I compiled after I patched
> it with mpls-linux-1.955 patch form sourceforge
>=20
> please help me to continue this work
>=20
>=20
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--=20
James R. Leu
jl...@mi...

[mpls-linux-general] vtysh not work

[<elc...@en...>]

Hello,

I installed the quagga package from sourceforge and then I typed the
command "vtysh" but this isn't work properly and I recieved this error:
vtysh: error while loading shared libraries: libtermcap.so.2: cannot open
shared object file : No such file or directory

for information:
I use the quagga.99.6 version
and the netkit uml kernel 2.6.18.8 debian which I compiled after I patched
it with mpls-linux-1.955 patch form sourceforge

please help me to continue this work

[mpls-linux-general] Kom je met een oldtimer, dan mag je gratis binnen!

[Christopher E. <pv...@51...>]

SREA UP Another 36.36%. Read This Hit List!

Score One Inc. (SREA)
Close: $0.60 UP 36.36%

In the last two days SREA has been on the watch list of OTCPicks.com,
OTCStockExchange.com, and Boonmarket.com rocketing it over 200%. Need we
say more? Get on SREA and ride the wave.

Locating the right alcohol rehab program for yourself or a loved one is
a tough, yet critical step on that path, and it is a step with which we
can help.
Na twee jaren sukkelen is eindelijk de attractie Vertigo geopend.
I forget about that sometime.

The substances are also more available in some places, and a really
strong reason is peer pressure.
To begin your search for an alcoholism treatment center that is right
for you or your loved one, simply click on the state below and a list of
cities in which we have listings will appear. I am beyond fed up and I
think I am angry enough to do something about it.

For the complete list of drugs of abuse, use the Drug Information link
to the left.

When you ask lawyers this question, you do not always get straight
answers because every circumstance is different. As an aside, I made one
embarrassing slip of the tongue during our conversation.

It may begin with coffee and cigarettes, then progress to marijuana or
amphetamines.

They are called functional alcoholics as they are very unaware of their
problem, having always been able to perform all the day-to-day
activities. She passed away, and I can never replace her. You, and I,
are two of millions.

com Psychiatric Hospital Specializing in the treatment of Women.
Victoria kraait en kirt In de Britse krant The Mirror stond dan weer dat
Victoria al terug in de studio zat en er van genoot opnieuw te zingen.
I'm pleased to present my interview, including the latest news about the
Million Cat Rescue, Randy's thoughts about American Idol, as well as
what's coming up professionally for Randy Jackson.

I'm pleased to present my interview, including the latest news about the
Million Cat Rescue, Randy's thoughts about American Idol, as well as
what's coming up professionally for Randy Jackson.

org introduces a new comprehensive blog updated daily by recovering
addicts and drug rehab professional.

Cocaine Addiction Help Starts Here.

Originally her name was going to be Vanilla Skye after some movie
character.

[mpls-linux-general] Now you will feel no cold, for each of you will be warmth to the other.

[Francis <kz...@jo...>]

Can't wait to see all the pics and I hope you two have a good and
relaxing honeymoon.
http://slmucus.hk/

Re: [mpls-linux-general] about ldpd.coonf

[srinath b. <sri...@ho...>]

Yes - I noticed that too during my LDP testing.
I usually enter the "mpls ldp" command by hand - this also works.



> Date: Thu, 21 Jun 2007 16:49:59 +0200> From: marco.schmidt@s2001.tu-chemn=
itz.de> To: ya...@gm...> CC: mpl...@li...>=
 Subject: Re: [mpls-linux-general] about ldpd.coonf> > Hello,> > the only c=
ommand you MUST not use is "mpls ip".> You have to configure it by hand in =
vtysh.> Have a look at a set of working config-files:> > http://www-user.tu=
-chemnitz.de/~smarco/MPLS/LDP/> > Another example is here:> > http://source=
forge.net/docman/display_doc.php?docid=3D49390&group_id=3D15443> > > Hello,=
> > > I have a problem with ldpd.> > When I launch ldpd, it reads the file =
ldpd.conf, gives the following message and stops> > ldpd: ldp_interface.c:5=
9: ldp_interface_create2: Assertion 'li->iff.index' failed. Aborted> > Actu=
ally the file is generated automatically by vtysh.> > If I leave the ldpd.c=
onf blank and use vtysh to configure ldpd it works fine.> > > Since I am wo=
rking on a project based on UML, typing the commands> > for each VM everyti=
me it starts is really a nightmare. > > Could anyone confirm if there is a =
bug in the program reading the> > ldpd.conf file? If not, may I have a corr=
ect example of ldpd.conf?> > Thanks in advance.> > > BTW, here is the gener=
ated ldpd.conf> > > !> > ! Zebra configuration saved from vty > > !   2007/=
06/20 05:00:08> > !> > hostname ldpd> > password zebra> > enable password z=
ebra> > !> > interface eth0> >  mpls ip> > !> > interface eth1> >  mpls ip>=
 > !> > interface gre0> > !> > interface lo> > !> > interface mpls0 > > !> =
> interface tunl0> > !> > !> > mpls ldp> >  egress connected> > !> > line v=
ty> > !> > > > -- > Marco Schmidt> mailto:marco.schmidt@s2001.tu-chemnitz.d=
e> > > > ------------------------------------------------------------------=
-------> This SF.net email is sponsored by DB2 Express> Download DB2 Expres=
s C - the FREE version of DB2 express and take> control of your XML. No lim=
its. Just data. Click to get it now.> http://sourceforge.net/powerbar/db2/>=
 _______________________________________________> mpls-linux-general mailin=
g list> mpl...@li...> https://lists.sourceforge=
.net/lists/listinfo/mpls-linux-general=

Re: [mpls-linux-general] about ldpd.coonf

[Marco S. <mar...@s2...>]

Hello,

the only command you MUST not use is "mpls ip".
You have to configure it by hand in vtysh.
Have a look at a set of working config-files:

http://www-user.tu-chemnitz.de/~smarco/MPLS/LDP/

Another example is here:

http://sourceforge.net/docman/display_doc.php?docid=3D49390&group_id=3D15443

> Hello,

> I have a problem with ldpd.
> When I launch ldpd, it reads the file ldpd.conf, gives the following mess=
age and stops
> ldpd: ldp_interface.c:59: ldp_interface_create2: Assertion 'li->iff.index=
' failed. Aborted
> Actually the file is generated automatically by vtysh.
> If I leave the ldpd.conf blank and use vtysh to configure ldpd it works f=
ine.

> Since I am working on a project based on UML, typing the commands
> for each VM everytime it starts is really a nightmare.=20
> Could anyone confirm if there is a bug in the program reading the
> ldpd.conf  file? If not, may I have a correct example of ldpd.conf?
> Thanks in advance.

> BTW, here is the generated ldpd.conf

> !
> ! Zebra configuration saved from vty=20
> !=A0=A0 2007/06/20 05:00:08
> !
> hostname ldpd
> password zebra
> enable password zebra
> !
> interface eth0
> =A0mpls ip
> !
> interface eth1
> =A0mpls ip
> !
> interface gre0
> !
> interface lo
> !
> interface mpls0=20
> !
> interface tunl0
> !
> !
> mpls ldp
> =A0egress connected
> !
> line vty
> !

>  =20
--=20
Marco Schmidt
mailto:mar...@s2...

[mpls-linux-general] about ldpd.coonf

[Xu Y. <ya...@gm...>]

Hello,

I have a problem with ldpd.
When I launch ldpd, it reads the file ldpd.conf, gives the following message
and stops
ldpd: ldp_interface.c:59: ldp_interface_create2: Assertion 'li->iff.index'
failed. Aborted
Actually the file is generated automatically by vtysh.
If I leave the ldpd.conf blank and use vtysh to configure ldpd it works
fine.

Since I am working on a project based on UML, typing the commands for each
VM everytime it starts is really a nightmare.
Could anyone confirm if there is a bug in the program reading the
ldpd.conffile? If not, may I have a correct example of
ldpd.conf?
Thanks in advance.

BTW, here is the generated ldpd.conf

!
! Zebra configuration saved from vty
!   2007/06/20 05:00:08
!
hostname ldpd
password zebra
enable password zebra
!
interface eth0
 mpls ip
!
interface eth1
 mpls ip
!
interface gre0
!
interface lo
!
interface mpls0
!
interface tunl0
!
!
mpls ldp
 egress connected
!
line vty
!

Re: [mpls-linux-general] Mpls-te in quagga

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

On Wed, Jun 20, 2007 at 09:17:38PM +0300, Dragoescu Alexandru wrote:
> Hy,
>=20
> Is mpls-te command in quagga compatible with actual linux-mpls
> implementation?
> Is there any way to make TE with mpls-linux using label distribution
> protocol ?

No.  the mpls-te stuff that comes in quagga is just to configure ospf to
send opaque LSAs with TE info.

> 10x,
> AlexD
> --=20
> Alexandru Dragoescu
> dra...@gm...
> QA engineer
> ITCNETWORKS Romania

> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general


--=20
James R. Leu
jl...@mi...

[mpls-linux-general] Mpls-te in quagga

[Dragoescu A. <dra...@gm...>]

Hy,

Is mpls-te command in quagga compatible with actual linux-mpls
implementation?
Is there any way to make TE with mpls-linux using label distribution
protocol ?


10x,
AlexD
-- 
Alexandru Dragoescu
dra...@gm...
QA engineer
ITCNETWORKS Romania

Re: [mpls-linux-general] Bandwidth reservation

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

I no longer support mpls-linux on the 2.4 kernels.  Also, the RSVP
implementation you are using is not part of mpls-linux.  Please see
the maintainer of that project for assistance.

On Wed, Jun 20, 2007 at 04:23:48PM +0530, Swapnil Bandekar wrote:
> Hi,
> i need help with getting bandwidth reservation working.
> i am using rsvpd with Linux kernel 2.4.18 with MPLS and HTB patched up.
> for label distribution and tunnel setup, i am using RSVPD with RSVP API=
=20
> version 5.02
>=20
> Thanks in advance
>=20
> --=20
> Swapnil Bandekar
> M. Tech. (Comm.)
> Electrical Engg. Dept.
> IIT Bombay
>=20
>=20
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--=20
James R. Leu
jl...@mi...

[mpls-linux-general] Bandwidth reservation

[Swapnil B. <swa...@ee...>]

Hi,
i need help with getting bandwidth reservation working.
i am using rsvpd with Linux kernel 2.4.18 with MPLS and HTB patched up.
for label distribution and tunnel setup, i am using RSVPD with RSVP API 
version 5.02

Thanks in advance

-- 
Swapnil Bandekar
M. Tech. (Comm.)
Electrical Engg. Dept.
IIT Bombay

Re: [mpls-linux-general] mpls ilm show behaviour

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

On Wed, Jun 20, 2007 at 10:24:04AM +1000, Michael Flower wrote:
> I have been running the comand 'mpls ilm show label xxx labelspace y'
> It always reports data on all the ilm's, not just for the specified
> label/labelspace.
>=20
> Is this the expected behaviour?

No.  It should filter them.  I'll look into fixing it.  Thank you for
reporting the bug.

> ...mcf
>=20
> --=20
>=20
>=20
> Michael Flower
> ES Labs
> Elantis Systems Inc.
> +61 3 9550 0866
>=20
>=20
> -------------------------------------------------------------------------
> This SF.net email is sponsored by DB2 Express
> Download DB2 Express C - the FREE version of DB2 express and take
> control of your XML. No limits. Just data. Click to get it now.
> http://sourceforge.net/powerbar/db2/
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--=20
James R. Leu
jl...@mi...

[mpls-linux-general] mpls ilm show behaviour

[Michael F. <mic...@el...>]

I have been running the comand 'mpls ilm show label xxx labelspace y'
It always reports data on all the ilm's, not just for the specified
label/labelspace.

Is this the expected behaviour?

...mcf

-- 


Michael Flower
ES Labs
Elantis Systems Inc.
+61 3 9550 0866