Hi,
After some start/stop testing, nfcapd began to refuse starting one of the two collector proccesses I was playing with (on different ports) with a message thus:
Mar 26 18:32:03 monitor /usr/local/bin/nfcapd[12288]: Another collector with different user ID is already running, and configured for '/var/flows/r1'
There were no nfcapd processes running.
I found this reference on the mailing list:
<http://www.mail-archive.com/nfsen-discuss@lists.sourceforge.net/msg00557.html>
Unfortunately, the bookkeeper.c I had was newer. So it seems that the bug is still there.
A reboot worked around the problem.
I modified the skeleton init script from Debian with this function:
do_start()
{
for Device in `ls /etc/nfcapd.d/*[!\~]` ; do
. $Device
start-stop-daemon --start --chuid nobody --pidfile $PIDFILE --exec $DAEMON -- $DAEMON_ARGS
done
}
The directory /etc/nfcapd.d contains files, one per device, with just a line like this:
DAEMON_ARGS="-w -D -l /var/flows/r1 -b 10.0.4.201 -p 10001"