2.4.6pre iptables masquerading seems to kill eth0

2.4.6pre iptables masquerading seems to kill eth0

[Helge Hafting]

I have a home network with two machines connected with
3c905B cards.  The main machine also has a isdn dialup connection.

Networking works well except if I let the main machine masquerade
so the other can use the internet too.  I use iptables for this.
It works for a day or so, then eth0 goes silent on the main machine.
(Rebooting it shows that the other one was fine all the time.)

The symptoms is that there is no contact between the two machines.
No ping or anything.  "ifconfig" shows the interface is up
with the correct ip address, but all packets just disappear.
There are no error messages except from programs that time out.

Bringing the interface down and up
again with ifconfig does not help.  It is compiled into the
kernel, so I can't try module reloading.

Is this some sort of known problem? Or is there something
I could do to find out more?  I couldn't
find anything in the logfiles.

Helge Hafting.

Re: 2.4.6pre iptables masquerading seems to kill eth0

[Thomas Weber]

I'm on 2.4.6pre3 + freeswan/ipsec on my gateway now for 5 days.
It's an old 486/66 32MB with several isdn links, a dsl uplink (with 
iptables masquerading) behind a ne2k clone and a 3c509 to the inside network.
no problems at all with the interfaces (all compiled as modules).

  Tom

In article <3B31A652.85D2E597@idb.hist.no>,
Helge Hafting <helgehaf@idb.hist.no> wrote:
>I have a home network with two machines connected with
>3c905B cards.  The main machine also has a isdn dialup connection.
>
>Networking works well except if I let the main machine masquerade
>so the other can use the internet too.  I use iptables for this.
>It works for a day or so, then eth0 goes silent on the main machine.
>(Rebooting it shows that the other one was fine all the time.)
>
>The symptoms is that there is no contact between the two machines.
>No ping or anything.  "ifconfig" shows the interface is up
>with the correct ip address, but all packets just disappear.
>There are no error messages except from programs that time out.
>
>Bringing the interface down and up
>again with ifconfig does not help.  It is compiled into the
>kernel, so I can't try module reloading.
>
>Is this some sort of known problem? Or is there something
>I could do to find out more?  I couldn't
>find anything in the logfiles.

Re: 2.4.6pre iptables masquerading seems to kill eth0

[Helge Hafting]

Thomas Weber wrote:
> 
> I'm on 2.4.6pre3 + freeswan/ipsec on my gateway now for 5 days.
> It's an old 486/66 32MB with several isdn links, a dsl uplink (with
> iptables masquerading) behind a ne2k clone and a 3c509 to the inside network.
> no problems at all with the interfaces (all compiled as modules).

Nice to know it works for you.  The troubled machine is a dual celeron, 
so it could be some sort of SMP problem.  I am trying pre5
to see if it is better, I'll probably know in a few days.

Helge Hafting

Re: 2.4.6pre iptables masquerading seems to kill eth0

[Thomas Weber]

On Fri, Jun 22, 2001 at 09:20:26AM +0200, Helge Hafting wrote:
> Thomas Weber wrote:
> > 
> > I'm on 2.4.6pre3 + freeswan/ipsec on my gateway now for 5 days.
> > It's an old 486/66 32MB with several isdn links, a dsl uplink (with
> > iptables masquerading) behind a ne2k clone and a 3c509 to the inside network.
> > no problems at all with the interfaces (all compiled as modules).
> 
> Nice to know it works for you.  The troubled machine is a dual celeron, 
> so it could be some sort of SMP problem.  I am trying pre5
> to see if it is better, I'll probably know in a few days.

since i just read another report on the list I checked again and it seems 
like i messed it up: 
you were talking about the 3c905, but I have a 3c509 (ISA) card.
I should have known, it's not the first time i mixed them up :(

  Tom