I have found several problems with the kismet init script. First, the method used to determine whether the wireless card has monitor mode doesn't work when using the wlan-ng drivers (wlan-ng doesn't have and 802.11b ioctls, so iwpriv doesn't give any useful information). I already spoke to latexer about this on irc (i'm Boogaloo). Second, and more importantly, the init script doesn't actually start kismet_server. The script reports that it was successful, but kismet_server is not running after the script completes and the pidfile doesn't exist. kismet_server will run is the --background option for start-stop-deamon is removed, but that's another problem for obvious reasons. Third, (latexer found this one, i'm just adding it here to remind him :) ) the pidfile created for kismet_hopper is different in start() and stop()....this is just a typo, really. One more thing i'd like to point out is that there's a line in /etc/kismet/kismet.conf that tells kismet_server what user to run as (for security purposes...the line is suiduser=user_name). kismet_server must run in a directory that this user has write permissions to (because it writes dump files, etc, to the working directory). Hope this is of some help, Justin
New tarball w/ initscripts should be propogating soon. Kismet server will _not_ start from the init script if you do not set a full path to a writeable directory for the logtemplate setting in /etc/kismet/kismet.conf e.g. suiduser=foo logtemplate=/home/foo/kismet/%n-%d-%i.%l Soon as i have a card to test the wlan-ng drivers w/ (coming in the mail tomorrow), i'll fix the script to work w/ wlan-ng stuff to. Until then, comment out the three lines in /etc/init.d/kismet
This bug is more than 180 days old. I'm resolving as TEST-REQUEST. Please reopen if the problem persists.
