First login?
Lost password?
 
 
sctp-discussion AT lists.franken.de 
SCTP Discussions

Text archives Help


Re: [SCTP-Discussion] Multihoming


Chronological Thread 
  • From: Michael Tuexen <Michael.Tuexen AT micmac.franken.de>
  • To: Patrick <stuedip AT student.ethz.ch>
  • Cc: <discussion AT sctp.de>
  • Subject: Re: [SCTP-Discussion] Multihoming
  • Date: Fri Oct 22 17:47:01 2004
  • List-id: SCTP Discussions <discussion.sctp.de>
Hi Patrik,

see my comments in-line.

Best regards
Michael

On Oct 22, 2004, at 14:18 Uhr, Stüdi, Patrick wrote:

Thank you Micheal,

Can you do the same with the combined_server and the terminal tool from
the sctplib?

I did and it works fine. shutting down one interface forces the stream to take the second path, as one would expect.

So it seems to be a problem with the sctpmultiserver/sctpterminal program. or my handling. I'm using sctlib-1.0.1 and socketapi-1.3.1.

You are using the socketapilib, so it might be an problem with your usage of
that library or a bug in the library.
Why does this happen? Have you closed the first association?
No, I'm not closing anyting. Actually, I double checked and in the two init packet are sent from the two interfaces, and the two init_acks as well. that's how it should be isn't it?

That is fine. You are using sctp_connectx, rigth?
Or are you using the the 1toMany style API and auto closing is happending?
I'm just using the example programs shipped with socketapi-1.3.1, I never heard about 1toMany style API, what is this?
It depends how you use SCTP. You can use it in a way similar to UDP (1toMany) or similar
to TCP (1to1). In the 1toMany style autoclosing happens if there is no traffic.

Do you still want to send DATA?
yes, and I'm expecting the data to go over the second interface when the primary is down. But from the ethereal dump I can see that the terminal program still tries to send over the primary. From time to time a FORWARD_TSN is sent as well, also on the primary, but of course none of these packets arrive at the server. Only a few Heartbeat packets sent over the second interface arrive at the server. And the second path is indeed up, I can ping the server on its second interface without any problem.
You should not see FORWARD_TSN cunks... That sounds like a bug.
We will release a new version of the sctplib and socketapilib
next week. So you might want to try again with that versions.

Best Regards,
Patrick






________________________________

From: Michael Tuexen
[mailto:Michael.Tuexen AT micmac.franken.de]
Sent: Wed 10/20/2004 9:06 PM
To: Stüdi, Patrick
Cc: discussion (E-mail)
Subject: Re: [SCTP-Discussion] Multihoming



See my comments below.

Best regards
Michael

On Oct 20, 2004, at 19:05 Uhr, Stüdi, Patrick wrote:

Hi,

thanks for the mail. I used ethereal now to trace the whole scenario.
remember I start sctpmultiserver on machineA (192.168.220.11,
192.168.221.12) and sctpterminal on machineB (192.168.220.7,
192.168.221.8).

are both addresses exchanged during the association setup?
yes, both addresses are exchanged in an init chunk (addresses of A)
and initACK (addresses of B).

then after exchanging an SCPT cookie packet, another init/initACK are
exchanged. then some
Why does this happen? Have you closed the first association? Or are you
using the
the 1toMany style API and auto closing is happending?
hearbeats/acks and finally data, transmitted from A's 192.168.220.11
to B's 192.169.220.7, which is the primary address. but after taking
interface eth1 down on machine B (which is 192.168.220.7) only
heartbeats are received by B's second interface, originating from A's
second interface.
Do you still want to send DATA?
Can you do the same with the combined_server and the terminal tool from
the sctplib and
run them with -V to see all notifications?

Thanks for helping.

Best regards,
Patrick



________________________________

From: Michael Tuexen
[mailto:Michael.Tuexen AT micmac.franken.de]
Sent: Tue 10/19/2004 7:21 PM
To: Stüdi, Patrick
Cc:
discussion AT sctp.de
Subject: Re: [SCTP-Discussion] Multihoming



Hi,

are both addresses exchanged during the association setup? Where
does the terminal client send the DATA packets? You could trace
the traffic on both interfaces to see what is happening. This
would help to understand what is going on.

Best regards
Michael

On Oct 19, 2004, at 14:16 Uhr, Stüdi, Patrick wrote:

Hi everyone,

I first want to apologize for my last post in case it sounded a bit
harsh. This is maybe because I'm somehow desperate with that
multihoming issue. Is there really no-one having an idea? Any help
would be appreciated. I have again attached my setup below.

Best Regards,

Patrick

----------------------------- SETUP
-----------------------------------------------------

I have two laptops, both with a WLAN card (which is configured to be
in 'ad hoc' mode) and a Bluetooth dongle.

I configured the network to looks as follows on machine A:



ifconfig

bnep0 Link encap:Ethernet HWaddr 00:10:DC:E9:6C:B8

inet addr:192.168.221.12 Bcast:192.168.221.255
Mask:255.255.255.0

UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1

RX packets:0 errors:0 dropped:0 overruns:0 frame:0

TX packets:0 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:100

RX bytes:116 (116.0 b) TX bytes:0 (0.0 b)



eth1 Link encap:Ethernet HWaddr 00:02:2D:7B:89:B4

inet addr:192.168.220.11 Bcast:192.168.220.255
Mask:255.255.255.0

UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1

RX packets:301 errors:0 dropped:0 overruns:0 frame:0

TX packets:57 errors:0 dropped:0 overruns:0 carrier:0

collisions:0 txqueuelen:100

RX bytes:45271 (44.2 Kb) TX bytes:9257 (9.0 Kb)

Interrupt:11 Base address:0x100



The configuration is the same on machine B. The route on A and B look
like:



Kernel IP routing table

Destination Gateway Genmask Flags Metric Ref
Use Iface

192.168.221.0 * 255.255.255.0 U 0 0
0 bnep0

192.168.220.0 * 255.255.255.0 U 0 0
0 eth1

127.0.0.0 * 255.0.0.0 U 0 0
0 lo



When testing, I can ping both interface without any problem. So I'm
starting the sctpmultiserver on machine B (the example program):



./sctpmultiserver



Right after, I start the terminal program on machine A to connect to
both interfaces.



./sctpterminal 192.168.220.7:echo 192.168.221.8:echo



Great. So echoing works fine. But now I'm taking down interface eth1.



ifdown eth1



The routing entry on machine B for eth1 disappears, on machine A the
table keeps consistent. But the example terminal program doesn't
switch to the secondary address. Interface two (192.168.221.8), is
pingable without any problems. I even tried to get rid of the 'eth1'
entry in A's routing table but this doesn't help either.

So, if anyone has a clue, please let me know. Any help would be
appreciated.

Thanks a lot.

Patrick

_______________________________________________
discussion mailing list
discussion AT sctp.de
http://www.sctp.de/mailman/listinfo/discussion




_______________________________________________
discussion mailing list
discussion AT sctp.de
http://www.sctp.de/mailman/listinfo/discussion




_______________________________________________
discussion mailing list
discussion AT sctp.de
http://www.sctp.de/mailman/listinfo/discussion




Archive powered by MHonArc 2.6.19+.

Top of page