Thank you Alexander! I did some further investigation on this case and I guess, I found out, what caused the error, which explains, why you were not able to reproduce the error. I logged messages to serial console and also connected a getty-terminal to it to observe the system state by running "top". I observed a drastic increase in memory usage, when reconnecting the fox board, so that at some point, there was no memory left at all and the system freezed. I noticed, that a lot of syslogd-processes were spawned (up to 5 instances), which were obviously responsible for consuming that much memory. What I did not mention explicitly was, that my sylogd daemon was configured to send all messages to any console, i.e. also to my telnet-session console. I sent quite a lot messages, and I did not terminate the telnet session before disconnecting the fox board. So when I reestablished the connection, the syslogd daemon seemed to try to send the messages, it was not able to send in disconnected state and that somehow lead to massive memory consumption and brought down the system. The problem does not occurr at all, if I close the telnet session before disconnecting the ethernet cable and the possiblity of error is dependent on the free memory available. I do not really understand, why syslogd behaves in such a way, when reconnecting (why it spawns several additional instances and why so much memory is consumed), perhaps someone knows an answer on this. It's actually no real problem to me anymore, as I can change the syslogd configuration, but it's still intersting to know, what's going on for informational purposes.
By the way, I do not use the classic red fox board, it's the newer (green) LX832. I am looking forward to any further insight into this topic! Thanks Dietmar
