[leaf-user] Bering 1.0 - stops responding during large upload

[Jonathan_Portwood <Jon...@ba...>]

I browsed the lists and net and couldn't find anyone with a similar problem
- 

I use Bering v1.0-Stable with kernel 2.4.20 & shorewall 1.3.12.

Whenever I try to upload large amounts of data (greater than 20 MB),
my lrp box fails.  Not sure of the problem. 

When it happens, I am unable to access the net from an internal client.
 My internal interface (eth1) is ok - everything seems ok from client
-> router. 
I can also run things like ntpdate and odsclient from my router box 
to net as well.  I just can't go from client -> net for some reason. 
I bring down my eth0 and am able to bring it back up and it grabs a valid
ip address.  I clear shorewall and restart it as well.  
To me, it seems client -> router is ok and i also think router -> net
is ok.  What is in between seems to be failing in this situation.  I could
find nothing interesting in the logs and ip -s link.  

I use RR cable service.  I can upload all day by bypassing the bering
box - so I don't believe it is RR or the modem.  

Rebooting clears the error - IF i restart my cable modem as well.  If
I reboot bering w/o restarting my modem, then I have to bring up my eth0
manually for some reason.

I run a small webserver with pictures - so if I am gone for the weekend
I notice when I come back everything's locked up.  Mainly a problem when
there is lots up upstream with very little or no downstream.  

Kind of bizaare.
Not sure if I explained it effectively - be happy  to clear something
up.

Any help would be appreciated.

Thanks.

Jonathan