I've also been seeing random crashes for the past month on CentOS 5.1
with v3.3 (ntop.x86_64 3.3-1.el5.rf). These started after we changed
the default settings for RRD and ntop itself. Most of the crashes leave
no information in /var/log/messages, only a segfault message from the
kernel.
Higher levels of network activity seem to make it crash more often (it
averages one crash every 6-12 hours. Sometimes the gap is as short as
45 minutes, other times it runs for over a day.
...
For the moment, we've resorted to running a script every 3 minutes via
cron to restart ntop when it falls over. It doesn't do anything to help
maintain our Global Traffic Statistics page totals, but it at least
keeps the Network Load graphs from having too many gaps.
#!/bin/bash
# Watches the ntop process
if [ ! `/usr/bin/pgrep -u ntop ntop` ]; then
echo "ntop is not running, restarting it"
/etc/init.d/ntop start;
else
echo "ntop is running"
fi
And we setup Nagios to keep an eye on things for times when the watch
script doesn't do its job.
_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop
