[LARTC] Question: TEQL via gateway

[LARTC] Question: TEQL via gateway

[Li, Ji]

Yesterday I posted a question, but I guess too much detail is provided
that no one would bother to read, so I rephrase the question and
hopefully some one may be willing to read. Sorry to spam.

I am using TEQL on two computers, each with two network interfaces. The
two eth0 are connected directly, and the two eth1 are connected via a
gateway (a linux machine). The problem is that when I send packets
through teql0 to the other machine (ping -I teql0), the packets can't go
through eth1 and all packets go via eth0. But I can send packets via
eth1 directly (ping -I eth1). What's wrong? 

P.S. I put eth1 and teql0 on one computer, and the gateway interface on
the same side into one subnet. I only found TEQL configuration for two
direct links. My routing tables are attached below.

Thanks a lot,
-Ji 


Node 1:
- eth0: 16.119.144.21
- eth1: 16.119.144.33
- teql0: 16.119.144.34

Destination     Gateway         Genmask         Flags Metric Ref    Use
Iface
16.119.144.66   *               255.255.255.255 UH    0      0        0
teql0
16.119.144.32   *               255.255.255.255 UH    0      0        0
eth1
16.119.144.64   16.119.144.32   255.255.255.252 UG    0      0        0
eth1
16.119.144.0    *               255.255.254.0   U     0      0        0
eth0
16.119.144.0    *               255.255.254.0   U     0      0        0
eth1
16.119.144.0    *               255.255.254.0   U     0      0        0
teql0
169.254.0.0     *               255.255.0.0     U     0      0        0
eth1
default         16.119.144.1    0.0.0.0         UG    0      0        0
eth0

Node 2:
- eth0: 16.119.144.27
- eth1: 16.119.144.65
- teql0: 16.119.144.66

Destination     Gateway         Genmask         Flags Metric Ref    Use
Iface
16.119.144.64   0.0.0.0         255.255.255.255 UH    0      0        0
eth1
16.119.144.34   0.0.0.0         255.255.255.255 UH    0      0        0
teql0
16.119.144.32   16.119.144.64   255.255.255.252 UG    0      0        0
eth1
16.119.144.0    0.0.0.0         255.255.254.0   U     0      0        0
eth0
16.119.144.0    0.0.0.0         255.255.254.0   U     0      0        0
eth1
16.119.144.0    0.0.0.0         255.255.254.0   U     0      0        0
teql0
169.254.0.0     0.0.0.0         255.255.0.0     U     0      0        0
eth1
0.0.0.0         16.119.144.1    0.0.0.0         UG    0      0        0
eth0

Gateway:
- eth0: 16.119.144.32
- eth1: 16.119.144.64

Destination     Gateway         Genmask         Flags Metric Ref    Use
Iface
16.119.144.64   0.0.0.0         255.255.255.252 U     0      0        0
eth1
16.119.144.32   0.0.0.0         255.255.255.252 U     0      0        0
eth0
16.119.144.0    0.0.0.0         255.255.254.0   U     0      0        0
eth1
16.119.144.0    0.0.0.0         255.255.254.0   U     0      0        0
eth1
127.0.0.0       0.0.0.0         255.0.0.0       U     0      0        0
lo
0.0.0.0         16.119.144.1    0.0.0.0         UG    0      0        0
eth0
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] Question: TEQL via gateway

[Taylor, Grant]

I think this has to do with an age old issue (problem or feature, you decide).  Linux (and most other unicies and other OSs in general) will pick the first interface on a subnet to use as the source interface to send the traffic out.  I think, this is more akin to what bonding is designed to answer.  I question are you wanting to really have the 2nd leg of the teql link be via a different path or were you just doing that to test to see if you were really sending traffic out both interfaces?  If you can have a direct cross over cable on both interfaces and you are wanting more aggregate bandwidth take a look at bonding, I think it will serve you better.  Disclaimer, I have not worked with teql or bonding my self directly and this is based on what reading I have done, that being more on bonding than teql.



Grant. . . .

Li, Ji wrote:
> Yesterday I posted a question, but I guess too much detail is provided
> that no one would bother to read, so I rephrase the question and
> hopefully some one may be willing to read. Sorry to spam.
> 
> I am using TEQL on two computers, each with two network interfaces. The
> two eth0 are connected directly, and the two eth1 are connected via a
> gateway (a linux machine). The problem is that when I send packets
> through teql0 to the other machine (ping -I teql0), the packets can't go
> through eth1 and all packets go via eth0. But I can send packets via
> eth1 directly (ping -I eth1). What's wrong? 
> 
> P.S. I put eth1 and teql0 on one computer, and the gateway interface on
> the same side into one subnet. I only found TEQL configuration for two
> direct links. My routing tables are attached below.
> 
> Thanks a lot,
> -Ji 
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

RE: [LARTC] Question: TEQL via gateway

[Li, Ji]

Yes, I really need two physical links to do my job. 
Also, when I connect two eth0 and eth1 directly, TEQL does send half of
the packets to each link. It only doesn't work when I put a gateway
between eth1. 
It seems that the way TEQL works is that it tries to put packets to both
interfaces. However, if one interface is not working, then it just sends
all packets to the other. When I "ping -I teql0 IP_of_teql0_of_the
other_end", eth1 broadcasts ARP for the MAC address of the destination,
and because the IP is in another subnet, it can't get any answer,
although I have specify the gateway for that IP. So TEQL sends all
packets to eth0. Don't know what's wrong.

-Ji 

-----Original Message-----
From: lartc-bounces@mailman.ds9a.nl
[mailto:lartc-bounces@mailman.ds9a.nl] On Behalf Of Taylor, Grant
Sent: Wednesday, June 22, 2005 3:01 PM
To: lartc@mailman.ds9a.nl
Subject: Re: [LARTC] Question: TEQL via gateway

I think this has to do with an age old issue (problem or feature, you
decide).  Linux (and most other unicies and other OSs in general) will
pick the first interface on a subnet to use as the source interface to
send the traffic out.  I think, this is more akin to what bonding is
designed to answer.  I question are you wanting to really have the 2nd
leg of the teql link be via a different path or were you just doing that
to test to see if you were really sending traffic out both interfaces?
If you can have a direct cross over cable on both interfaces and you are
wanting more aggregate bandwidth take a look at bonding, I think it will
serve you better.  Disclaimer, I have not worked with teql or bonding my
self directly and this is based on what reading I have done, that being
more on bonding than teql.



Grant. . . .

Li, Ji wrote:
> Yesterday I posted a question, but I guess too much detail is provided

> that no one would bother to read, so I rephrase the question and 
> hopefully some one may be willing to read. Sorry to spam.
> 
> I am using TEQL on two computers, each with two network interfaces. 
> The two eth0 are connected directly, and the two eth1 are connected 
> via a gateway (a linux machine). The problem is that when I send 
> packets through teql0 to the other machine (ping -I teql0), the 
> packets can't go through eth1 and all packets go via eth0. But I can 
> send packets via
> eth1 directly (ping -I eth1). What's wrong? 
> 
> P.S. I put eth1 and teql0 on one computer, and the gateway interface 
> on the same side into one subnet. I only found TEQL configuration for 
> two direct links. My routing tables are attached below.
> 
> Thanks a lot,
> -Ji
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc