I have several different daemons running on a server. I want to write a sure fire script for cron to make sure the daemon is still running (every 10 minutes). I am getting close but I can't figure out how to finish the script. I tried checking for the existance of a pid file but it is there and the daemon isn't running.

I can check manually check for it several ways (ps or netstat). Below is the general syntax I'm trying:

if [ ! netstat -nl | grep :137 ]; then
/usr/local/samba/sbin/nmbd
fi

It is checking to see if the server is listening on any port 137. If it isn't then the [ ] is null and should execute the line below it. I keep getting an error:
./check.samba: [: missing `]'
grep: ]: No such file or directory

I tried to do this also: `netstat -nl | grep :137` but I get different errors. It seems I just don't know what syntax to use.

Any help??

Google/see Freshmeat for "monit". Much easier, dependable.
And if you don't want Monit to die, start it from init (after appropriate testing).

I found a temporary solution (maybe a permanent one). Now I have it in cron to check every 10 minutes to see if the daemonname is running. Here is the syntax I used:

if [ -z `ps -ef |grep daemonname |grep -v grep | awk '{print$2}'` ]; then
/path/to/start/daemon
fi

There were 2 things I was doing wrong: needed ` ` to enclose the expression and I was using -n which checks to see if an "expression exist AND if is NOT NULL". The -z checks to see if the "expression exist and is NULL". Those 2 parts were causing the errors. http://www.freeos.com/guides/lsst/ was very helpful in getting this done.

I plan on looking at monit to see if it does what I need without having to spend too much time getting a PHD in it.

Thanks for the help.

Yeah, I sometimes forget to address the actual question...
Here's a oneliner you can run from cron:
"pgrep <daemonname> 2>&1>/dev/null || /path/to/start/<daemonname>"
Note this doesn't work if crond goes down for any reason.