[Q] Weird network-crashes with DFE-500TX

[Peter Mogensen]

Hi,

  I'm getting desperate so I have a question. I'm not a TPC/IP
  or Ethernet wizard so this may sound stupid.

  We have had a RedHat 5.1 (2.0.35) machine with a DFE-500TX
  and tulip 0.88 running for months without problems. Then suddenly
  on day it starts to loose network connection randomly once or twice
  every day. Uptime range from 2 days to 2 hours.
  We re-installed the system with RH5.2 (2.0.36) / tulip 0.89H
  and changed the net card (another DFE-500TX). The problem remained.
  Now we've changed to an NE2000/ISA card and the problem seems to
  have disappeared.
  The only event which I can relate to the start of our troubles is
  that out ISP changed the router which we use as gateway the same day.
  The new router is a small Cisco-thing.

  So the question is:

  Is it possible that there's a bug in the tulip driver which causes
  it to crash when talking to a Cisco router?

  We have no traces in the log-files prior to any crash and the happen
  very randomly.
  Our default gateway is on eth0:1 so it's not a standard setup. That
  might be a factor.
  The server does not crash. You can log in on the console and:

  /etc/rc.d/init.d/network stop
  rmmod tulip
  insmod tulip
  /etc/rc.d/init.d/network start

  ... and the system runs again (for a while)

  Is it possible that it's related to the Cisco router or should we
  look for a periodic hardware error?

  thanks,

       Peter