Robert <(E-Mail Removed)> wrote:
> Robert wrote:
[network stops sporadically]
>> My network keeps stopping after some time, around a
>> day or something. I'll have to walk to my server and do
>> something to get it working again. I've disabled apmd
>> and didn't compile apm into the kernel. My cable between
>> those computers is ok so i don't have a clue on what it
>> could be. It worked for about 60 days until after some
>> time my little brother resetted the server. I don't really
>> want to reinstall Red Hat 9 if i don't to but if nobody
>> knows a solution...
>>
>> I'm using my server for different purposes, it runs a dhcp
>> server, a ftp server, a http server, a ssh server and a mysql
>> server.
>>
>> It is also used for masquerading on two interfaces.
>> eth0 = connected to ethernet adsl modem
>> eth1 (192.168.0.1) = connected to my pc (192.168.0.2)
>> eth2 (192.168.1.2) = connected to my parents pc (192.168.1.2)
>>
>> I don't know if the connection fails to all network interfaces
>> but i don't think so. I'll try my parents computer
>> for some time to see if it fails there too.
Make more clear what "connection fails" mean. Is this
"only" the masqueraded Interface connection or also
*any* attempt to communicate with the server?
> I just had time to test if it only occurs with eth1 because
> it happened again
and the result is that it only happens
> with eth1, the original settings for this card in
> redhat-config-network aren't changed. The network reactivates
> as soon as i ping my computer from the server, plz help.
Check Link/Activity-LEDs and error counters viewed by
ifconfig on both ends of the ethernet link.
Try to reproduce the error without waiting so long:
- Check whether the problem is ARP related. Clear ARP
cache on your PC and then try to ping the server.
- Move interface eth0 on your PC down and up again.
Does the interface get operational again?
Check log file whether your DHCP-Server works properly.
If in doubt disable DHCP for eth1 on server and eth0
on your PC, and use static configuration.
Check other log files on server and your PC whether
there are suspicous entries around the time of failure.
Exchange eth1 and eth2 NICs and watch whether the error
sticks with the NIC. (I got strange errors with a NIC
which sporadically failed; it still indicated with LEDs
and error counters that everything is fine, though).
If the error does not stick with the NIC, swapp cables
also between eth1 and eth2. However I don't think that
this will cause the error but it doesn't hurt to check
it.
When you get the failure, use tcpdump to sniff eth1 on
server and eth0 on your PC simultaneously. Then try
to ping from your PC the server, after this ping from
the server your PC.
If everything fails, you could use a wrong-around and
start on the server a periodically ping job pinging
your PC. Yuck, Yuck!
HTH
Ciao, Horst
--
»When pings go wrong (It hurts me too)« E.Clapton/E.James/P.Tscharn