Skip to contentSkip to conversation

TEQL questions

LARTC Chris Murray 3 participants 5 messages Feb 20, 2002 – Feb 22, 2002

lore
Share & tools
lore
#0 lore raw

Hi All -

I've decided that TEQL will meet my needs. I am having a problem with it now and I have not been able to find any documentation in regards to this.

I have added both of my interfaces (on both ends of the connections) to the teql0 interface:

tc qdisc add dev eth1 root teql0
tc qdisc add dev eth2 root teql0

I did this for both ends and assigned an ip address to teql0 on both ends of 172.16.3.1/29 and 172.16.3.2/29 respectivly.

ip addr add dev teql0 172.16.3.1/29
ip addr add dev teql0 172.16.3.2/29

I cannot get them to ping, I get Network is unreachable.

I am I missing something ?

I have the ip addresses of 172.16.2.1/29 + 172.16.1.2/29 on one box and 172.16.2.2/29 + 172.16.1.1/29 on the other machine. These links can see each other fine.

Thanks in advance for any help.

- Chris

--
[email protected]

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/

#1 lore raw

I never tried it myself, but I found this in the kernel sources :

Short description:
------------------

                  +-------+   eth1   +-------+
                  |       |==========|       |
  'network 1' ----|   A   |          |   B   |---- 'network 2'
                  |       |==========|       |
                  +-------+   eth2   +-------+

Router A and B are connected to each other with two links. The device to be set up is basically a roundrobbin distributor over eth1 and eth2, for sending packets. No data ever comes in over an teql device, that just appears on the 'raw' eth1 and eth2. But now we just have devices, we also need proper routing. One way to do this is to assign a /31 network to both links, and a /31 to the teql0 device as well. That means, we have to have a transfer network for this.

Routing: eth1_A: 10.0.0.0/31

             eth1_B: 10.0.0.1/31

             eth2_A: 10.0.0.2/31
             eth2_B: 10.0.0.3/31

             teql0_A: 10.0.0.4/31
             teql0_B: 10.0.0.5/31

             If you only have Linux boxes on each side, you can use /31
             as subnet and so you can avoid wasting IP-Addresses for
             broadcast and network addresses

You have to set up the IP-Addresses accordingly. Your default gw should point to teql0.

In order to use this feature TEQL must be enables in your kernel. You have to set up this framework on both sides (ie. Router A and B).

CAVEATS:
--------

You have to disable Reverse Path Filtering! Ie. I guess you do not want to use this in a Firewall/DMZ area. :-)
Then there is the nasty problem of packet reordering. Let's say 6 packets need to be sent from A to B - eth1 might get 1, 3 and 5. eth2 would then do 2, 4 and 6. In an ideal world, router B would receive this in order, 1, 2, 3, 4, 5, 6. But the possibility is very real that the kernel gets it like this: 2, 1, 4, 3, 6, 5. The problem is that this confuses TCP/IP. While not a problem for links carrying many different TCP/IP sessions, you won't be able to to a bundle multiple links and get to ftp a single file lots faster, except when your receiving or sending OS is Linux, which is not easily shaken by some simple reordering.

Stef

Quoted 30 lines · you

Hi All -

I've decided that TEQL will meet my needs. I am having a problem with it now and I have not been able to find any documentation in regards to this.

I have added both of my interfaces (on both ends of the connections) to the teql0 interface:

tc qdisc add dev eth1 root teql0
tc qdisc add dev eth2 root teql0

I did this for both ends and assigned an ip address to teql0 on both ends of 172.16.3.1/29 and 172.16.3.2/29 respectivly.

ip addr add dev teql0 172.16.3.1/29
ip addr add dev teql0 172.16.3.2/29

I cannot get them to ping, I get Network is unreachable.

I am I missing something ?

I have the ip addresses of 172.16.2.1/29 + 172.16.1.2/29 on one box and 172.16.2.2/29 + 172.16.1.1/29 on the other machine. These links can see each other fine.

Thanks in advance for any help.

- Chris

Signature · 6 lines
[email protected]
 More QOS info : http://www.docum.org/
 Title : "Using Linux as bandwidth manager"
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/
#2 lore raw
In reply to Chris Murray

If figured out why it didn't work after following the howto, there was no mention is that you have to do a:

ip link set dev teql0 up

On both sides of the link. I am now starting my testing with two 1.5mbps links, and it seems to be configured correctly But now it appears that all my traffic is traveling over eth2 and not equally divided between eth1 and eth2. I tested by opening two ftp sessions from the remote machine to and ftp server on the internet and I only got 1/2 of my expected bandwith

Thanks - Chris

Signature · 7 lines
Chris Murray                    Network Services Specialist
[email protected]             Stargate Connections, Inc.
http://www.stargate.ca/         ph. +1 (604) 606-8988

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/
#3 lore raw
In reply to Chris Murray
Quoted 4 lines · Chris Murray

If figured out why it didn't work after following the howto, there was no mention is that you have to do a:

ip link set dev teql0 up

ah! Fixed & added you to the 'thanks to' list.

Quoted 5 lines

On both sides of the link. I am now starting my testing with two 1.5mbps links, and it seems to be configured correctly But now it appears that all my traffic is traveling over eth2 and not equally divided between eth1 and eth2. I tested by opening two ftp sessions from the remote machine to and ftp server on the internet and I only got 1/2 of my expected bandwith

Have you tcpdumped to verify this assertion? There are issues with packet reordering which look just like this, but do not indicate that teql is not working. So, could you run tcpdump on teql0, on eth1 and eth2, and check where traffic is going?

Thanks!

Regards,

bert

Signature · 7 lines
http://www.PowerDNS.com          Versatile DNS Software & Services
http://www.tk                              the dot in .tk
Netherlabs BV / Rent-a-Nerd.nl           - Nerd Available -
Linux Advanced Routing & Traffic Control: http://ds9a.nl/lartc
_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/
#4 lore raw
In reply to Chris Murray
Quoted 3 lines · bert hubert
Quoted 1 line

ip link set dev teql0 up

ah! Fixed & added you to the 'thanks to' list.

Cool. Thanks!

Quoted 4 lines

Have you tcpdumped to verify this assertion? There are issues with packet reordering which look just like this, but do not indicate that teql is not working. So, could you run tcpdump on teql0, on eth1 and eth2, and check where traffic is going?

I've done some tests and I can see traffic on teql0 and eth2, but I don't see any related traffic on eth1. I tried initiating several downloads from various webservers and all traffic appeared on eth2. eth1 just shows me the arps on my network, as does eth2.

Any things to check?

Thanks - Chris

Signature · 7 lines
Chris Murray                    Network Services Specialist
[email protected]             Stargate Connections, Inc.
http://www.stargate.ca/         ph. +1 (604) 606-8988

_______________________________________________
LARTC mailing list / [email protected]
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/lartc/