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

Re: [mpls-linux-general] MPLS+IPv6

[Bruno J. M. <bj...@me...>]

Hi!

Are there any changes regarding iptables IPv6 structures manipulation?

Can You help me on that?

Bruno

On Fri, 2 Apr 2004, James R. Leu wrote:

> It could very well be that I'm not modifing the correct structures for
> IPv6. I will look into that.
>
> On Tue, Mar 30, 2004 at 08:04:48PM +0100, Bruno Jorge Marcelino wrote:
> >
> > I think the problem is in iproute patch. Maybe it's only working with
> > IPv4 kernel routing structures??
> >
> > I looked at the code, and seems that route table manipulation is done
> > by the same manner in ldp-portable (and this only works in IPv4)
> >
> > I established an mpls route (IPv4) between 2 routers and ping was
> > successful!
> >
> > 	ip route add 10.0.0.1/32 dev eth1 spec_nh 0x8842 0x2
> >
> > I'm right or I'm completely wrong?!?! :)
> >
> > Thank's
> > Bruno
> >
> > On Fri, 26 Mar 2004, James R. Leu wrote:
> >
> > > On Fri, Mar 26, 2004 at 06:04:41PM +0000, Bruno Jorge Marcelino wrote:
> > > >
> > > > Yes. I applied the mpls patch in the kernel and build it.
> > >
> > > I will create a patch that adds debugging to the IPv6 stack to se can
> > > see what is happening.
> > >
> > > > On Fri, 26 Mar 2004, James R. Leu wrote:
> > > >
> > > > > Did you build the kernel fro a patch?  It so, I can send you a patch that
> > > > > adds some debugging.  Otherwise I will have to build a custom RPM for you.
> > > > >
> > > > > On Fri, Mar 26, 2004 at 03:47:52PM +0000, Bruno Jorge Marcelino wrote:
> > > > > >
> > > > > > I've checked dmesg and seems the packets aren't entering the MPLS stack...
> > > > > >
> > > > > > The only messages that I have are from (modified) mplsadm2 commands:
> > > > > > 	mplsadm2 -d
> > > > > > 	mplsadm2 -A -O 0 -o push.gen.56.set.eth1.ipv6.$NET::65
> > > > > >
> > > > > > MPLSADM mplsadm6.c:696:main: Out segment input: 0
> > > > > > MPLSADM mplsadm6.c:692:main: Out instr input:
> > > > > > push.gen.56.set.eth1.ipv6.fc97:86d5:b2ed::65
> > > > > > Key: 0x00000002
> > > > > > Out Segment add: Success
> > > > > > MPLSADM mplsadm6.c:296:fill_instructions: Instruction: push
> > > > > > MPLSADM mplsadm6.c:233:fill_label: Label type: gen
> > > > > > MPLSADM mplsadm6.c:296:fill_instructions: Instruction: set
> > > > > > MPLSADM mplsadm6.c:185:parse_nh_info: eth1
> > > > > > MPLSADM mplsadm6.c:186:parse_nh_info: fc97:86d5:b2ed::65
> > > > > > MPLSADM mplsadm6.c:208:parse_nh_info: Nexthop protocol: ipv6
> > > > > > MPLSADM mplsadm6.c:600:fill_instructions: Length: 2
> > > > > > Out Instr: Success
> > > > > >
> > > > > > 	output of dmesg:
> > > > > >
> > > > > > MPLS DEBUG net/mpls/mpls_ioctls.c:109:mpls_ioctl: IOCTL Add NHLFE 35847
> > > > > > MPLS DEBUG net/mpls/mpls_out_info.c:322:mpls_add_out_label: enter
> > > > > > MPLS DEBUG net/mpls/mpls_out_info.c:395:mpls_add_out_label: added MOI-NHLFE moi-2
> > > > > > MPLS DEBUG net/mpls/mpls_out_info.c:396:mpls_add_out_label: exit
> > > > > > MPLS DEBUG net/mpls/mpls_ioctls.c:180:mpls_ioctl: IOCTL Set OUT Instr. 35864
> > > > > > MPLS DEBUG include/net/mpls.h:665:mpls_out_info_hold: moi-2-get 3
> > > > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:203:mpls_build_opcode_push: enter
> > > > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:216:mpls_build_opcode_push: exit
> > > > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:853:mpls_build_opcode_set: enter
> > > > > > MPLS DEBUG net/mpls/mpls_dst.c:170:mpls_dst_alloc: enter
> > > > > > MPLS DEBUG net/mpls/mpls_dst.c:229:mpls_dst_alloc: exit(c1663bc0)
> > > > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:918:mpls_build_opcode_set: exit
> > > > > > MPLS DEBUG net/mpls/mpls_utils.c:349:mpls_instr_register: enter
> > > > > > MPLS DEBUG net/mpls/mpls_utils.c:352:mpls_instr_register: registering OP-00-PUSH
> > > > > > MPLS DEBUG net/mpls/mpls_utils.c:356:mpls_instr_register: exit
> > > > > > MPLS DEBUG net/mpls/mpls_utils.c:349:mpls_instr_register: enter
> > > > > > MPLS DEBUG net/mpls/mpls_utils.c:352:mpls_instr_register: registering OP-01-SET
> > > > > > MPLS DEBUG net/mpls/mpls_utils.c:356:mpls_instr_register: exit
> > > > > >
> > > > > > routerPC5:/usr/src/mpls-linux/utils# cat /proc/net/mpls/out
> > > > > > MOI Key:0x00000002 Instruction Set
> > > > > > PUSH PUSH(gen 56)
> > > > > > SET SET(eth1,0.0.0.0)
> > > > > >
> > > > > > I first used
> > > > > > 	ip6tables -A OUTPUT -d $NET:2::/64 -j spec_nh --spec_nh 0x8847:0x2
> > > > > > and then
> > > > > > 	ip -f inet6 route add $NET:2::/64 dev eth1 spec_nh 0x8847 0x2
> > > > > >
> > > > > > and none of the above commands worked! The packets aren't using MPLS to leave the router...
> > > > > >
> > > > > > I have an IPv6 network with 5 PC's running debian testing/unstable, kernel
> > > > > > 2.6.2+mpls-1.922.
> > > > > >
> > > > > > Can you help me resolve my problem/improving IPv6 support in
> > > > > > MPLS and ldp-portable?
> > > > > >
> > > > > > I can make some changes in the code, but I you need to point me on what to do!
> > > > > >
> > > > > > Regarding the changes in mplsadm2, you can see the parsing of commands
> > > > > > has changed from ":" to "." (because in IPv6, separator is now ":"). This
> > > > > > conflicts with IPv4 addresses...
> > > > > >
> > > > > > Bruno
> > > > > >
> > > > > >
> > > > > > On Wed, 24 Mar 2004, James R. Leu wrote:
> > > > > >
> > > > > > > On Wed, Mar 24, 2004 at 08:26:38PM +0000, Bruno Jorge Marcelino wrote:
> > > > > > > > Hi there!
> > > > > > > >
> > > > > > > > I'm doing some research on MPLS and IPv6 for my master thesis...
> > > > > > > >
> > > > > > > > Right now, i'm using kernel 2.6.2 with mpls-1.922 patch!
> > > > > > >
> > > > > > > I will be releasing 1.930 soon, with support for MPLS ICMP and better
> > > > > > > internal locking and referencing.
> > > > > > >
> > > > > > > > I've changed mplsadm2 to work with IPv6 address, and by looking at
> > > > > > > > debugging messages, I think that's all OK (no errors are found).
> > > > > > >
> > > > > > > Please send me a patch, I'd like to include support for IPv6.
> > > > > > >
> > > > > > > > Exception is in /proc/net/mpls/in, where  SET(ethX,X.X.X.X) appears
> > > > > > > > with 0.0.0.0 (because macro NIPQUAD(a) only works with IPv4 addresses, right?).
> > > > > > >
> > > > > > > This should be an easy to fix.
> > > > > > >
> > > > > > > > But I have a problem when binding FEC to LSP (with mpls patched
> > > > > > > > ip6tables): it isn't working! The traffic is not leaving the router using
> > > > > > > > mpls.
> > > > > > > >
> > > > > > > > Do you have any idea on the problem?
> > > > > > >
> > > > > > > Did you check the MPLS kernel debug?  (execute the command 'dmesg').
> > > > > > > Did the packets ever enter the MPLS stack?  What command did you
> > > > > > > use?  None of the IPv6 code has been tested.  I do not have an IPv6 network
> > > > > > > to test with (nor do I know how to set one up).
> > > > > > >
> > > > > > > Have you tried 'ip'?
> > > > > > >
> > > > > > > ip -f inet6 route add <ipv6 route>/<mask len> dev eth0 spec_nh 0x8847 <key>
> > > > > > >
> > > > > > > > What's the status on porting ldp-portable to 2.6 kernel with IPv6?
> > > > > > >
> > > > > > > ldp-portable has the structures to handle IPv6, but no code has been written
> > > > > > > to handle IPv6.  I would be glad to point you to the places I know that
> > > > > > > would be affected by IPv6.
> > > > > > >
> > > > > > > --
> > > > > > > James R. Leu
> > > > > > > jl...@mi...
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > -------------------------------------------------------
> > > > > > This SF.Net email is sponsored by: IBM Linux Tutorials
> > > > > > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > > > > > GenToo technologies. Learn everything from fundamentals to system
> > > > > > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > > > > > _______________________________________________
> > > > > > mpls-linux-general mailing list
> > > > > > mpl...@li...
> > > > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > >
> > > > > --
> > > > > James R. Leu
> > > > > jl...@mi...
> > > > >
> > > >
> > > >
> > > >
> > > > -------------------------------------------------------
> > > > This SF.Net email is sponsored by: IBM Linux Tutorials
> > > > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > > > GenToo technologies. Learn everything from fundamentals to system
> > > > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > > > _______________________________________________
> > > > mpls-linux-general mailing list
> > > > mpl...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > >
> > > --
> > > James R. Leu
> > > jl...@mi...
> > >
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
> --
> James R. Leu
> jl...@mi...
>

[mpls-linux-general] Regarding iptables

[priya S. <mpl...@ya...>]

Haii
 
I have setup an mpls network and i am trying to do traffic engineering...for which i need to mark packets....
 
i was trying to execute the iptables command...
 
iptables -A OUTPUT -t mangle -p tcp -j FTOS --set-ftos 0x28
 
but it displays an error saying
iptables: No chain/target/match by that name..
 
 
I downloaded the iptabes-1.2.4-0.2-dscp.tgz frm http://dsmpls.atlantis.ugent.be/
 
I applied the dscp patch by making patch-o-matic....but d'nt get to see DSCP patch....so d'nt apply it.....
 
Wht is the reson for this error 
 

Thanks & Regards
Sirisha




---------------------------------
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th

[mpls-linux-general] problem setting up 2 LSP's simultaneously

[priya S. <mpl...@ya...>]

Hi,
 
I have a network set up as shown...
                                                         Eth2

                                                       |-------------Core Router 2
                eth1     eth1         eth2      |                     eth1

                                                       |                    |
Edge Router1 ---- Core Router 1---------|                    |

                                        Eth0        |                 Eth1

                                                       |------------ Core Router 3------------------Edge Router2

                                                           Eth0                       eth3             eth1

 

 

I want to set up two LSP's 

1) ER1 -CR1- CR2- CR3- ER2

2) ER1-CR1-CR3-ER2

 

 

I used 

 

#vi lspsetup

10.0.7.2 10.0.1.1 200 0 10.0.1.2:10.0.3.2:0 1

10.0.7.2 10.0.1.1 210 0 10.0.1.2:10.0.2.2:10.0.4.1:0 1

 

#rtest2 -f lspsetup

 

 

to set up the lsp's. when I do them individually it works fine but when i try to set them up simulateneously and run 

 

# tunnel  -L -c 

LSPID          Destination (type label/ phb /          viface)            Packets            Bytes

E 0                                 (gen 21680/ BE/          T21680eth1)          0                      0

 

At ER2

 

# ./rapirecv_auto 

Building LSP from 10.0.1.1:210 to 10.0.7.2 LSPD 200 EXT LSPID 0.0.0.0

Rapi_Session: sid =19,fd=4 0

Filterspecs 2 Flowspecs 0

rapi_reserve(): 0

other or RESV without PATH

 

i also tried setting up two LSP's on a single path....which also yielded the same results.

 

kindly let us know what I have to do in order to have two LSP's set up simultaneously

 

thanks in advance

 

Sirisha




---------------------------------
Do you Yahoo!?
Yahoo! Tax Center - File online by April 15th

[mpls-linux-general] Little Bro Make U Shy To Screw?? Hehe paretic nullius

[Inez E. <udi...@cc...>]

barf firstclass stizidae secondbest distance bolivia brunt 
nessun airpipe parenthood susurrate songstress workman alnico ripe mapping esker mien 
chemise lackadaisy sleazy dumfound solitaire fattish beadle dimetrodon paddy abuse loudmouth morose quackery appalachia sluggish exhort 

Re: [mpls-linux-general] Regd routing daemons and OSPF

[Ramon C. <cas...@in...>]

On Thu, 8 Apr 2004, kavitha ganapa wrote:

> Hi Ramon,
>
> I did take a look at the OSPF TE RFCs and other related RFCs. I have also
> gone through the code for OSPFAPI  and could get the LSA header information
> through the ospfapiclient. But the problem is the ospf daemons on each of

I'm having difficulties to understand your problem: you have a OSPF
network, that works fine without MPLS. You state that you can get the
LSA header: which ones ? all the area/network ?





> the routers are communicating but not the routers themselves. That is I give

This is what I don't understand. The OSPF daemons establish an adjacency
between them, they exchange the link state databae, but you cannot ping
them?

* Where does MPLS fit in all this? Are you trying to establish an
adjacency between routers using an LSP as a Forwarding Adjacency?
(although you should not do that)

* Which LSPs do you want to establish ?

* How are you classifying/filtering the traffic? Which ip rules did you
enable ?


Can you describe your topology a little more ? and what you plan to do ? I
cannot fit this with your original problem of calculating paths from link
state info.



> a telnet or ping from one to another, there is no response. Basically, I
> have mpls enabled linux kernel rpm installed, source code for zebra with the

Did you make sure that you compiled zebra using the kernel headers that
were patched for MPLS ?


> ospfapi. I have tried with ldp and rsvp as signaling protocols. But in vain.

Can you at least set a LSP manually between routers ?



> When I installed the zebra-ldp rpm and ran the opsfd it pings fine. When the
> ospfd from the source (zebra with ospfapiclient) is running why doesn't it
> ping.


Withour more info, I would say that maybe you did not compile the last
vesion of zebra with the right ldp patch and using the patched kernel
headers.... Did you have a reject when patching?


Best regards,


Ramon.

Re: [mpls-linux-general] Regd routing daemons and OSPF

[kavitha g. <kav...@ho...>]

Hi Ramon,

I did take a look at the OSPF TE RFCs and other related RFCs. I have also 
gone through the code for OSPFAPI  and could get the LSA header information 
through the ospfapiclient. But the problem is the ospf daemons on each of 
the routers are communicating but not the routers themselves. That is I give 
a telnet or ping from one to another, there is no response. Basically, I 
have mpls enabled linux kernel rpm installed, source code for zebra with the 
ospfapi. I have tried with ldp and rsvp as signaling protocols. But in vain. 
When I installed the zebra-ldp rpm and ran the opsfd it pings fine. When the 
ospfd from the source (zebra with ospfapiclient) is running why doesn't it 
ping.

Please let me know,
I am all stuck,
Thanks a lot,
Kavitha


>>Date: Fri, 26 Mar 2004 07:21:05 +0100 (MET)
>>From: Ramon Casellas <cas...@in...>
>>X-X-Sender: cas...@me...
>>To: kavitha ganapa <kg...@en...>
>>Cc: mpl...@li...
>>Subject: Re: [mpls-linux-general] Regd routing daemons and OSPF
>>X-Filter-Version: 1.14 (green.eng.utoledo.edu)
>>X-Status:
>>X-Keywords:
>>
>>On Thu, 25 Mar 2004, kavitha ganapa wrote:
>>
>> >
>> > Hi All,
>> >
>> > We have OSPF daemon running on the MPLS network to find just the 
>>shortest
>> > path from one node to another. What if I want to find different set of
>> > paths than what OSPF finds. How do I go about it?
>>
>>
>>Hi Kavitha,
>>
>>You could start taking a look at OSPF TE extensions (which basically add
>>opaque LSAs) to redistribute link state information like max bandwidth and
>>then use OSPFAPI (included in the software distribution) to perform the
>>path computation. This means hat you need to code in C, and that you will
>>reuse an OSPF library, but it should work. I plan to work on this soon, so
>>keep in touvh with the list :)
>>
>>Regards,
>>Ramon
>
>

_________________________________________________________________
Persistent heartburn? Check out Digestive Health & Wellness for information 
and advice. http://gerd.msn.com/default.asp

AW: [mpls-linux-general] kernel 2.4

[Zordan A. <zor...@gm...>]

Hi, 

> What distribution I should take for building/patching 
> 2.4.(21|25) kernel?

I've set up a MPLS environment on a RedHat 9.0 and it
works well except the ldp-portable....
i have the same issue like you ...
Zebra and ospf are running and distributing their
routings tables, but when i started mplsd the
routers cannot establish a ldp session.

Regards

Z.A.

-- 
+++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz

[mpls-linux-general] kernel 2.4

[Roman E. <rr...@wn...>]

	Hello!

What distribution I should take for building/patching 2.4.(21|25) kernel?

Previous attempt (about a year ago) of building MPLS enabled network 
was unsuccesfull. Zebras didn't want to establish LDP adjacencies one 
with another. I hope something got changed and want try to setup it again
without changing kernel version.

-- 
R:Em

Re: [mpls-linux-general] (no subject)

[Muhammad R. S. <sa...@cc...>]

thank you very much....another question is that is it necessary to use =
the TTL bits within the MPLS header...I would  like to utilize those 8 =
bits in encoding some other information.
regards
  ----- Original Message -----=20
  From: James R. Leu=20
  To: Muhammad R. Sami=20
  Cc: mpl...@li...=20
  Sent: Friday, April 02, 2004 8:35 PM
  Subject: Re: [mpls-linux-general] (no subject)


  I see no reason why you couldn't.

  On Fri, Apr 02, 2004 at 08:09:48PM +0300, Muhammad R. Sami wrote:
  > so this means that I can stream video over MPLS network assuming =
that RTP will take care of the realtime issues of such traffic.=20
  >   ----- Original Message -----=20
  >   From: James R. Leu=20
  >   To: Muhammad R. Sami=20
  >   Cc: mpl...@li...=20
  >   Sent: Friday, April 02, 2004 7:51 PM
  >   Subject: Re: [mpls-linux-general] (no subject)
  >=20
  >=20
  >   I'm not sure what you mean.
  >=20
  >   RTP and RTCP are just protocols with run on top of IP.  All IP =
traffic can
  >   run on top of MPLS.  So I guess the answer to your question is =
yes.
  >=20
  >   On Fri, Apr 02, 2004 at 07:28:14PM +0300, Muhammad R. Sami wrote:
  >   > Does the MPLS implementation support RTP and RTCP protocols for =
video streaming?=20
  >   > regards
  >   --=20
  >   James R. Leu
  >   jl...@mi...
  >=20
  >=20
  >   -------------------------------------------------------
  >   This SF.Net email is sponsored by: IBM Linux Tutorials
  >   Free Linux tutorial presented by Daniel Robbins, President and CEO =
of
  >   GenToo technologies. Learn everything from fundamentals to system
  >   =
administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=3Dcli=
ck
  >   _______________________________________________
  >   mpls-linux-general mailing list
  >   mpl...@li...
  >   https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

Re: [mpls-linux-general] (no subject)

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

I see no reason why you couldn't.

On Fri, Apr 02, 2004 at 08:09:48PM +0300, Muhammad R. Sami wrote:
> so this means that I can stream video over MPLS network assuming that RTP will take care of the realtime issues of such traffic. 
>   ----- Original Message ----- 
>   From: James R. Leu 
>   To: Muhammad R. Sami 
>   Cc: mpl...@li... 
>   Sent: Friday, April 02, 2004 7:51 PM
>   Subject: Re: [mpls-linux-general] (no subject)
> 
> 
>   I'm not sure what you mean.
> 
>   RTP and RTCP are just protocols with run on top of IP.  All IP traffic can
>   run on top of MPLS.  So I guess the answer to your question is yes.
> 
>   On Fri, Apr 02, 2004 at 07:28:14PM +0300, Muhammad R. Sami wrote:
>   > Does the MPLS implementation support RTP and RTCP protocols for video streaming? 
>   > regards
>   -- 
>   James R. Leu
>   jl...@mi...
> 
> 
>   -------------------------------------------------------
>   This SF.Net email is sponsored by: IBM Linux Tutorials
>   Free Linux tutorial presented by Daniel Robbins, President and CEO of
>   GenToo technologies. Learn everything from fundamentals to system
>   administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
>   _______________________________________________
>   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] (no subject)

[Muhammad R. S. <sa...@cc...>]

so this means that I can stream video over MPLS network assuming that =
RTP will take care of the realtime issues of such traffic.=20
  ----- Original Message -----=20
  From: James R. Leu=20
  To: Muhammad R. Sami=20
  Cc: mpl...@li...=20
  Sent: Friday, April 02, 2004 7:51 PM
  Subject: Re: [mpls-linux-general] (no subject)


  I'm not sure what you mean.

  RTP and RTCP are just protocols with run on top of IP.  All IP traffic =
can
  run on top of MPLS.  So I guess the answer to your question is yes.

  On Fri, Apr 02, 2004 at 07:28:14PM +0300, Muhammad R. Sami wrote:
  > Does the MPLS implementation support RTP and RTCP protocols for =
video streaming?=20
  > regards
  --=20
  James R. Leu
  jl...@mi...


  -------------------------------------------------------
  This SF.Net email is sponsored by: IBM Linux Tutorials
  Free Linux tutorial presented by Daniel Robbins, President and CEO of
  GenToo technologies. Learn everything from fundamentals to system
  =
administration.http://ads.osdn.com/?ad_id=3D1470&alloc_id=3D3638&op=3Dcli=
ck
  _______________________________________________
  mpls-linux-general mailing list
  mpl...@li...
  https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

Re: [mpls-linux-general] (no subject)

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

I'm not sure what you mean.

RTP and RTCP are just protocols with run on top of IP.  All IP traffic can
run on top of MPLS.  So I guess the answer to your question is yes.

On Fri, Apr 02, 2004 at 07:28:14PM +0300, Muhammad R. Sami wrote:
> Does the MPLS implementation support RTP and RTCP protocols for video streaming? 
> regards
-- 
James R. Leu
jl...@mi...

[mpls-linux-general] (no subject)

[Muhammad R. S. <sa...@cc...>]

Does the MPLS implementation support RTP and RTCP protocols for video =
streaming?=20
regards

AW: [mpls-linux-general] Segmentation fault

[Zordan A. <zor...@gm...>]

Hi again,

i have now a running zebra with ospf ... and i can even start mpsld without
a segmentation fault 
(the problem is based in the zebra.conf and the ospfd.conf. if they are not
correct, mpls will not work ... in my case: i don't set the directly
connected nics into the same subnet)

Now the routers try to set up a ldp session, but i got
ldp_session_attempt_setup: MPLS_NON_BLOCKING
and ldp_session_attempt_setup: MPLS_FAILURE messages. (both routers are
sending ldp-hello messages)
A further problem is the crash of mplsd each time i made a show ldp neighbor
in the mplsd-console ...

I looked in the online repository for a newer version of ldp, but i only
found the 0.310 which i already have.

Is this one of the ldp-bugs or does i forget anything ? (i have no time to
wait on the next release, 'cause i should set up the mpls-environment in the
next days)

Regards 

Zordan Arres

-- 
+++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz

Re: [mpls-linux-general] Segmentation fault @Richard Doerfler

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

You are using a very old version of zebra.  Most of the bugs you will
encounter have been fixed, but a new release has not been made.  If you
are interested in assisting in developing ldp-portable please see
http://mpls-linux.sf.net/ to see how to get the most recent version from
my repository.  Otherwise if you just want to use zebra+ldp for testing
or learning, I suggest waiting until a new release is made.

On Wed, Mar 31, 2004 at 11:07:21AM +0200, Zordan Arres wrote:
> Hi,
> 
> I've just installed ldp and try a very simple configuration, shown as follow
> 
> --------------------------- zebra config ------------------------------
> hostname LER1
> password root
> enable password root
> service advanced-vty
> !
> interface lo
>  description Loopback 
> !
> interface dummy0
>  ip address 192.168.13.5/32
> !
> interface eth0
>  ip address 192.168.13.1/32
> !
> interface eth1
>  ip address 192.168.13.2/32
> !
> line vty
>  exec-timeout 0 0
> 
> --------------------------- ospfd config ------------------------------
> 
> hostname LER1-ospf
> password root
> enable password root
> !
> interface lo
> !
> interface teql0
> !
> interface eth0
> !
> interface eth1
> !
> interface dummy0
> !
> router ospf
>  passive-interface lo
>  network 192.168.13.0/24 area 0
>  network 192.168.0.1/32 area 0
> !
> line vty
>  exec-timeout 0 0
> 
> --------------------------- mplsd config ------------------------------
> 
> hostname LER1-mpls
> password root
> enable password root
> !
> mpls ldp
> !
> interface lo
> !
> interface eth0
>  mpls ldp
> !
> interface eth1
>  mpls ldp
> !
> interface dummy0
> !
> line vty
>  exec-timeout 0 0
> 
> After that i run zebra, ospfd and mplsd, but the mplsd command exists
> With a segmentation error ...
> 
> I use the 2.4 kernel and zebra+ldp 0.94.
> 
> As i saw, you had the same problem. How did you fix it ?
> 
> Thanks for your help
> 
> Zordan Arres 
> 
> -- 
> +++ NEU bei GMX und erstmalig in Deutschland: T?V-gepr?fter Virenschutz +++
> 100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

[mpls-linux-general] Segmentation fault @Richard Doerfler

[Zordan A. <zor...@gm...>]

Hi,

I've just installed ldp and try a very simple configuration, shown as follow

--------------------------- zebra config ------------------------------
hostname LER1
password root
enable password root
service advanced-vty
!
interface lo
 description Loopback 
!
interface dummy0
 ip address 192.168.13.5/32
!
interface eth0
 ip address 192.168.13.1/32
!
interface eth1
 ip address 192.168.13.2/32
!
line vty
 exec-timeout 0 0

--------------------------- ospfd config ------------------------------

hostname LER1-ospf
password root
enable password root
!
interface lo
!
interface teql0
!
interface eth0
!
interface eth1
!
interface dummy0
!
router ospf
 passive-interface lo
 network 192.168.13.0/24 area 0
 network 192.168.0.1/32 area 0
!
line vty
 exec-timeout 0 0

--------------------------- mplsd config ------------------------------

hostname LER1-mpls
password root
enable password root
!
mpls ldp
!
interface lo
!
interface eth0
 mpls ldp
!
interface eth1
 mpls ldp
!
interface dummy0
!
line vty
 exec-timeout 0 0

After that i run zebra, ospfd and mplsd, but the mplsd command exists
With a segmentation error ...

I use the 2.4 kernel and zebra+ldp 0.94.

As i saw, you had the same problem. How did you fix it ?

Thanks for your help

Zordan Arres 

-- 
+++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz

Re: [mpls-linux-general] MPLS+IPv6

[Bruno J. M. <bj...@me...>]

I think the problem is in iproute patch. Maybe it's only working with
IPv4 kernel routing structures??

I looked at the code, and seems that route table manipulation is done
by the same manner in ldp-portable (and this only works in IPv4)

I established an mpls route (IPv4) between 2 routers and ping was
successful!

	ip route add 10.0.0.1/32 dev eth1 spec_nh 0x8842 0x2

I'm right or I'm completely wrong?!?! :)

Thank's
Bruno

On Fri, 26 Mar 2004, James R. Leu wrote:

> On Fri, Mar 26, 2004 at 06:04:41PM +0000, Bruno Jorge Marcelino wrote:
> >
> > Yes. I applied the mpls patch in the kernel and build it.
>
> I will create a patch that adds debugging to the IPv6 stack to se can
> see what is happening.
>
> > On Fri, 26 Mar 2004, James R. Leu wrote:
> >
> > > Did you build the kernel fro a patch?  It so, I can send you a patch that
> > > adds some debugging.  Otherwise I will have to build a custom RPM for you.
> > >
> > > On Fri, Mar 26, 2004 at 03:47:52PM +0000, Bruno Jorge Marcelino wrote:
> > > >
> > > > I've checked dmesg and seems the packets aren't entering the MPLS stack...
> > > >
> > > > The only messages that I have are from (modified) mplsadm2 commands:
> > > > 	mplsadm2 -d
> > > > 	mplsadm2 -A -O 0 -o push.gen.56.set.eth1.ipv6.$NET::65
> > > >
> > > > MPLSADM mplsadm6.c:696:main: Out segment input: 0
> > > > MPLSADM mplsadm6.c:692:main: Out instr input:
> > > > push.gen.56.set.eth1.ipv6.fc97:86d5:b2ed::65
> > > > Key: 0x00000002
> > > > Out Segment add: Success
> > > > MPLSADM mplsadm6.c:296:fill_instructions: Instruction: push
> > > > MPLSADM mplsadm6.c:233:fill_label: Label type: gen
> > > > MPLSADM mplsadm6.c:296:fill_instructions: Instruction: set
> > > > MPLSADM mplsadm6.c:185:parse_nh_info: eth1
> > > > MPLSADM mplsadm6.c:186:parse_nh_info: fc97:86d5:b2ed::65
> > > > MPLSADM mplsadm6.c:208:parse_nh_info: Nexthop protocol: ipv6
> > > > MPLSADM mplsadm6.c:600:fill_instructions: Length: 2
> > > > Out Instr: Success
> > > >
> > > > 	output of dmesg:
> > > >
> > > > MPLS DEBUG net/mpls/mpls_ioctls.c:109:mpls_ioctl: IOCTL Add NHLFE 35847
> > > > MPLS DEBUG net/mpls/mpls_out_info.c:322:mpls_add_out_label: enter
> > > > MPLS DEBUG net/mpls/mpls_out_info.c:395:mpls_add_out_label: added MOI-NHLFE moi-2
> > > > MPLS DEBUG net/mpls/mpls_out_info.c:396:mpls_add_out_label: exit
> > > > MPLS DEBUG net/mpls/mpls_ioctls.c:180:mpls_ioctl: IOCTL Set OUT Instr. 35864
> > > > MPLS DEBUG include/net/mpls.h:665:mpls_out_info_hold: moi-2-get 3
> > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:203:mpls_build_opcode_push: enter
> > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:216:mpls_build_opcode_push: exit
> > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:853:mpls_build_opcode_set: enter
> > > > MPLS DEBUG net/mpls/mpls_dst.c:170:mpls_dst_alloc: enter
> > > > MPLS DEBUG net/mpls/mpls_dst.c:229:mpls_dst_alloc: exit(c1663bc0)
> > > > MPLS DEBUG net/mpls/mpls_opcode_all.c:918:mpls_build_opcode_set: exit
> > > > MPLS DEBUG net/mpls/mpls_utils.c:349:mpls_instr_register: enter
> > > > MPLS DEBUG net/mpls/mpls_utils.c:352:mpls_instr_register: registering OP-00-PUSH
> > > > MPLS DEBUG net/mpls/mpls_utils.c:356:mpls_instr_register: exit
> > > > MPLS DEBUG net/mpls/mpls_utils.c:349:mpls_instr_register: enter
> > > > MPLS DEBUG net/mpls/mpls_utils.c:352:mpls_instr_register: registering OP-01-SET
> > > > MPLS DEBUG net/mpls/mpls_utils.c:356:mpls_instr_register: exit
> > > >
> > > > routerPC5:/usr/src/mpls-linux/utils# cat /proc/net/mpls/out
> > > > MOI Key:0x00000002 Instruction Set
> > > > PUSH PUSH(gen 56)
> > > > SET SET(eth1,0.0.0.0)
> > > >
> > > > I first used
> > > > 	ip6tables -A OUTPUT -d $NET:2::/64 -j spec_nh --spec_nh 0x8847:0x2
> > > > and then
> > > > 	ip -f inet6 route add $NET:2::/64 dev eth1 spec_nh 0x8847 0x2
> > > >
> > > > and none of the above commands worked! The packets aren't using MPLS to leave the router...
> > > >
> > > > I have an IPv6 network with 5 PC's running debian testing/unstable, kernel
> > > > 2.6.2+mpls-1.922.
> > > >
> > > > Can you help me resolve my problem/improving IPv6 support in
> > > > MPLS and ldp-portable?
> > > >
> > > > I can make some changes in the code, but I you need to point me on what to do!
> > > >
> > > > Regarding the changes in mplsadm2, you can see the parsing of commands
> > > > has changed from ":" to "." (because in IPv6, separator is now ":"). This
> > > > conflicts with IPv4 addresses...
> > > >
> > > > Bruno
> > > >
> > > >
> > > > On Wed, 24 Mar 2004, James R. Leu wrote:
> > > >
> > > > > On Wed, Mar 24, 2004 at 08:26:38PM +0000, Bruno Jorge Marcelino wrote:
> > > > > > Hi there!
> > > > > >
> > > > > > I'm doing some research on MPLS and IPv6 for my master thesis...
> > > > > >
> > > > > > Right now, i'm using kernel 2.6.2 with mpls-1.922 patch!
> > > > >
> > > > > I will be releasing 1.930 soon, with support for MPLS ICMP and better
> > > > > internal locking and referencing.
> > > > >
> > > > > > I've changed mplsadm2 to work with IPv6 address, and by looking at
> > > > > > debugging messages, I think that's all OK (no errors are found).
> > > > >
> > > > > Please send me a patch, I'd like to include support for IPv6.
> > > > >
> > > > > > Exception is in /proc/net/mpls/in, where  SET(ethX,X.X.X.X) appears
> > > > > > with 0.0.0.0 (because macro NIPQUAD(a) only works with IPv4 addresses, right?).
> > > > >
> > > > > This should be an easy to fix.
> > > > >
> > > > > > But I have a problem when binding FEC to LSP (with mpls patched
> > > > > > ip6tables): it isn't working! The traffic is not leaving the router using
> > > > > > mpls.
> > > > > >
> > > > > > Do you have any idea on the problem?
> > > > >
> > > > > Did you check the MPLS kernel debug?  (execute the command 'dmesg').
> > > > > Did the packets ever enter the MPLS stack?  What command did you
> > > > > use?  None of the IPv6 code has been tested.  I do not have an IPv6 network
> > > > > to test with (nor do I know how to set one up).
> > > > >
> > > > > Have you tried 'ip'?
> > > > >
> > > > > ip -f inet6 route add <ipv6 route>/<mask len> dev eth0 spec_nh 0x8847 <key>
> > > > >
> > > > > > What's the status on porting ldp-portable to 2.6 kernel with IPv6?
> > > > >
> > > > > ldp-portable has the structures to handle IPv6, but no code has been written
> > > > > to handle IPv6.  I would be glad to point you to the places I know that
> > > > > would be affected by IPv6.
> > > > >
> > > > > --
> > > > > James R. Leu
> > > > > jl...@mi...
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > > -------------------------------------------------------
> > > > This SF.Net email is sponsored by: IBM Linux Tutorials
> > > > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > > > GenToo technologies. Learn everything from fundamentals to system
> > > > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > > > _______________________________________________
> > > > mpls-linux-general mailing list
> > > > mpl...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > >
> > > --
> > > James R. Leu
> > > jl...@mi...
> > >
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > 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] iptables traffic eng.

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

Use a MPLS enabled iproute2.

ip route add 1.1.1.1/32 via 192.168.1.1 lsp 0x2

On Tue, Mar 30, 2004 at 09:33:47AM +0200, Zordan Arres wrote:
> > It sounds to me like you want seom sort of transparent L2 
> > connection, like ethernet over MPLS.  I have implemented such 
> > a beast in for the 2.4 kernel.  It shouldn't be too hard to 
> > port it to 2.6.
> 
> ... like this ... I wan't to try out the practical use of mpls in the
> Core net of a routing network such as an isp use it ... It simply
> Should route the traffic of the connected hosts in the right way and if
> I figure out how it works, even route different classes of traffic
> In different ways or to handle the traffic of an "premium" customer with
> A higher priority. Actually i only know, how to decide the route of
> A packet through the net with iptables and that doesn't work correctly....
>  (as i wrote earlier ...)
> 
> Actually i use the Kernel 2.4, 'cause some network drivers didn't work
> With 2.6 ... 
> So, how can i use your implementation (which params do i have to use ?) ?
> 
> -- 
> +++ NEU bei GMX und erstmalig in Deutschland: T?V-gepr?fter Virenschutz +++
> 100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

AW: [mpls-linux-general] iptables traffic eng.

[Zordan A. <zor...@gm...>]

> It sounds to me like you want seom sort of transparent L2 
> connection, like ethernet over MPLS.  I have implemented such 
> a beast in for the 2.4 kernel.  It shouldn't be too hard to 
> port it to 2.6.

... like this ... I wan't to try out the practical use of mpls in the
Core net of a routing network such as an isp use it ... It simply
Should route the traffic of the connected hosts in the right way and if
I figure out how it works, even route different classes of traffic
In different ways or to handle the traffic of an "premium" customer with
A higher priority. Actually i only know, how to decide the route of
A packet through the net with iptables and that doesn't work correctly....
 (as i wrote earlier ...)

Actually i use the Kernel 2.4, 'cause some network drivers didn't work
With 2.6 ... 
So, how can i use your implementation (which params do i have to use ?) ?

-- 
+++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz

Re: [mpls-linux-general] FEC binding for src and dst address

[Rudy S W. <rs...@an...>]

I'm trying to associate a label and some QoS parameters  with a
specific FEC the uses both the source and destination IP address (IP
packet from the ingress LER). I want to give different QoS service levels
to IP hosts connected to an ingress LER that may be destined for the same
egress LER interface.

   IP(10.2.2.2)
     \
      ingress LER -------- egress LER----IP(10.4.4.5)
     /
   IP(10.1.1.1)

If traffic from 10.1.1.1 and 10.2.2.2 want to reach 10.4.4.5, they can
have two different labels.  In the label mapping procedure, I want to be
able to send two different IP address in the FEC tlv with one associated
with the source IP address and another with the destination IP address. Is
there any mechanism in ldp-portable to accomplish this?

~rsw

On Mon, 29 Mar 2004, James R. Leu wrote:

> What are you trying to accomplish, there might be a better way to
> do this.  For example you could use the path tlv to know where a
> label request originated (source).
>
> On Mon, Mar 29, 2004 at 05:45:50PM -0500, Rudy S Wen wrote:
> > Hello,
> >
> > I've been working with ldp-portable for the past few months in order to
> > adopt it to a thesis project that I'm completing.  I'm wondering if there
> > is support to negotiate a label binding for an FEC that uses both a source
> > and destination IP address? From what I've seen in ldp_nortel.h, and fec
> > can only be associated with either a source or destination IP addr but
> > not both. If i were to modify the ldp-portable to support this type of
> > FEC, are there any major difficulties that may occur? I know I've got to
> > modify ldp_fec.c, ldp_state_func.c, mpls_global.h, and ldp_nortel.c to
> > accomodate this change, would this be all that is necessary?
> >
> > Thanks for any help!
> > ~rsw
> >
> >
> >
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > 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] FEC binding for src and dst address

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

What are you trying to accomplish, there might be a better way to
do this.  For example you could use the path tlv to know where a
label request originated (source).

On Mon, Mar 29, 2004 at 05:45:50PM -0500, Rudy S Wen wrote:
> Hello,
> 
> I've been working with ldp-portable for the past few months in order to
> adopt it to a thesis project that I'm completing.  I'm wondering if there
> is support to negotiate a label binding for an FEC that uses both a source
> and destination IP address? From what I've seen in ldp_nortel.h, and fec
> can only be associated with either a source or destination IP addr but
> not both. If i were to modify the ldp-portable to support this type of
> FEC, are there any major difficulties that may occur? I know I've got to
> modify ldp_fec.c, ldp_state_func.c, mpls_global.h, and ldp_nortel.c to
> accomodate this change, would this be all that is necessary?
> 
> Thanks for any help!
> ~rsw
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

[mpls-linux-general] FEC binding for src and dst address

[Rudy S W. <rs...@an...>]

Hello,

I've been working with ldp-portable for the past few months in order to
adopt it to a thesis project that I'm completing.  I'm wondering if there
is support to negotiate a label binding for an FEC that uses both a source
and destination IP address? From what I've seen in ldp_nortel.h, and fec
can only be associated with either a source or destination IP addr but
not both. If i were to modify the ldp-portable to support this type of
FEC, are there any major difficulties that may occur? I know I've got to
modify ldp_fec.c, ldp_state_func.c, mpls_global.h, and ldp_nortel.c to
accomodate this change, would this be all that is necessary?

Thanks for any help!
~rsw

Re: [mpls-linux-general] iptables traffic eng.

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

It sounds to me like you want seom sort of transparent L2 connection,
like ethernet over MPLS.  I have implemented such a beast in for the 2.4
kernel.  It shouldn't be too hard to port it to 2.6.

On Mon, Mar 29, 2004 at 04:30:02PM +0200, Zordan Arres wrote:
> > Why do you want to use FORWARD entries?  I've only tried using
> > MPLS + iptables in OUTPUT entries.  FORWARD entries haven't gone through
> > IPv4 output processing and are not ready to be hadned off to another
> > layer (ie MPLS).  Do things work as expected if you remove the FORWARD
> > entries?
> 
> i've set up a mpls-routing net ... each edge router
> is a gateway for one of the 192.168*** net ... now i wan't to
> "tunnel" the mpls-net to connect to the net of the
> second ler. this doesn't work without forwarding-entrys, 'cause
> otherwise the arp request isn't labled and cannot be routed
> correctly through the network !
> 
> e.g. i have a computer with the ip 192.168.0.1 connected on ler1
> this one should connect to 192.168.1.1 on ler2. To set up the
> connection the 0.1 send a packet with the arp who-has request to ler1 ...
> without any forwarding entry this router sends the packet unlabled into
> the network and because he doesn't have a direct connection to
> 1.1 he returned no arp-reply. If i entered an iptables-entry as described
> earlier he labled simply each incoming packet from 0.1 and route it through
> the net, so it can be recieved by 1.1 ... 
> 
> -- 
> +++ NEU bei GMX und erstmalig in Deutschland: T?V-gepr?fter Virenschutz +++
> 100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> 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] iptables traffic eng.

[Zordan A. <zor...@gm...>]

> Why do you want to use FORWARD entries?  I've only tried using
> MPLS + iptables in OUTPUT entries.  FORWARD entries haven't gone through
> IPv4 output processing and are not ready to be hadned off to another
> layer (ie MPLS).  Do things work as expected if you remove the FORWARD
> entries?

i've set up a mpls-routing net ... each edge router
is a gateway for one of the 192.168*** net ... now i wan't to
"tunnel" the mpls-net to connect to the net of the
second ler. this doesn't work without forwarding-entrys, 'cause
otherwise the arp request isn't labled and cannot be routed
correctly through the network !

e.g. i have a computer with the ip 192.168.0.1 connected on ler1
this one should connect to 192.168.1.1 on ler2. To set up the
connection the 0.1 send a packet with the arp who-has request to ler1 ...
without any forwarding entry this router sends the packet unlabled into
the network and because he doesn't have a direct connection to
1.1 he returned no arp-reply. If i entered an iptables-entry as described
earlier he labled simply each incoming packet from 0.1 and route it through
the net, so it can be recieved by 1.1 ... 

-- 
+++ NEU bei GMX und erstmalig in Deutschland: TÜV-geprüfter Virenschutz +++
100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz

Re: [mpls-linux-general] iptables traffic eng.

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

Why do you want to use FORWARD entries?  I've only tried using
MPLS + iptables in OUTPUT entries.  FORWARD entries haven't gone through
IPv4 output processing and are not ready to be hadned off to another
layer (ie MPLS).  Do things work as expected if you remove the FORWARD
entries?

On Mon, Mar 29, 2004 at 03:27:03PM +0200, Zordan Arres wrote:
> Hi, 
>  
> i've set up a little MPLS network under rh9. 
> I use 4 pcs in the Core network and 2 Edge Routers to 
> bind 2 different networks together: one with 192.168.0.0/24 
> and the other 192.168.1.0/24.  
> now i wan't to divide different classes of traffic. e.g.  
> route ssh over one edge of the core net and everything else over another. 
>  
> i tried to use iptables with the following commands: 
> iptables -t mangle -A FORWARD -p tcp --dport 22 -j MPLS --set-key 0x2 
> iptables -t mangle -A OUTPUT -p tcp --dport 22 -j MPLS --set-key 0x2 
>  
> and 
> iptables -t mangle -A FORWARD -d 192.168.1.0/24 -j MPLS --set-key 0x3 
> iptables -t mangle -A OUTPUT -d 192.168.1.0/24 -j MPLS --set-key 0x3 
>  
> but every time i changed the FORWARD-Entries the LER alway ignored my
> changes. 
> The only thing that worked is to reboot them ... is this a bug or did i 
> forget something ?? 
>  
> regards 
>  
> Zordan Arres 
> 
> -- 
> +++ NEU bei GMX und erstmalig in Deutschland: T?V-gepr?fter Virenschutz +++
> 100% Virenerkennung nach Wildlist. Infos: http://www.gmx.net/virenschutz
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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