eth1 breaks eth0

I have a debian install on a machine with 2 eth cards,eth0 and eth1. eth0 has been configured using pppoeconf for a pppoe connection which is working fine. eth1 was meant to be there in order to provide for a nat so that an internet connection can be provided to another computer also using debian. The moment I plug in the cable to eth1 , eth0 stops working. (Is this caused maybe by hotplug ?) The goal was to actually get eth1 working also without breaking eth0's connection and afterwards make a nat using iptables. The same thing was done using Slackware and netconfig+pppoeconf and there were no problems. The pppoe is up using the pppd service and it was configured using pppoeconf.

I tried to find out the problem by trying to find out which file is beeing modified when I plug the cable in eth1 so I did strace ifconfig eth1 2>&1 | grep open I got a list of files and I suspect /proc/net/if_inet6 is the file that ifconfig modifies and reads thata from. So I decided to do sudo lsof | grep if_inet6 in order to find out what process uses the hotplug feature to autodetect when I put the cable inside eth1. Unfortunately I find no process that did this.

What should I do so that the internet connection doesn't drop when I plug in the cable for eth1 ?

NOTE: Posting from groups.google.com (or some web-forums) dramatically reduces the chance of your post being seen. Find a real news server.

You're posting from a search engine - and maybe you'd have better luck by first using it to locate debian specific _newsgroups_ and if not successful, trying the Usenet newsgroup 'comp.os.linux.networking'.

OK - _before_ and after plugging in the cable, run the commands

/sbin/ifconfig -a /sbin/route -n

Most likely.

Very good!! Most people wouldn't think of that.

Are you actually _using_ IPv6? This would create a ifconfig output that contains a line saying something like

eth0 Link encap:Ethernet HWaddr 00:60:97:32:4A:6A inet addr:192.0.2.11 Bcast:192.0.2.255 Mask:255.255.255.0 inet6 addr: fe80::260:97ff:fe32:4a6a/64 Scope:Link

and the key is that third line. If the address begins with 'f' (here 'fe80:') IPv6 probably isn't doing you any good. You seem to be posting from a Romanian IP range, and there are only six valid IPv6 ranges used in that country:

2001:0b30::/32 2001:1518::/32 2001:4098::/32 2001:4b48::/32 2001:4d80::/32 2a01:07d8::/32

Only if the 'inet6 addr:' begins with one of these eight digits (the /32 means the netmask is 32 bits with - eight digits) do you have a working IPv6 address. In spite of the fact that IPv6 has been around for 13 years, and that classic IPv4 addresses (such as 192.0.2.11) are in very short supply, not many ISPs are supporting it.

Can't help there - I don't use 'hotplug'.

Try posting to comp.os.linux.networking - mentioning the version of Debian you are using (etch, lenny, sid), and show the output of /sbin/ifconfig and /sbin/route before and after plugging in. Not many people in _this_ Usenet newsgroup know about *nix.

Old guy