Automating the "killing" process via a shell script

fatrandy13 · 10-06-2005, 10:07 AM

when I log into the UNIX server at work, it loads up twm by defualt, and the very first thing i do every morning is kill twm and run startkde. I want to automate this....

I started a shell script, but seem to be stuck..

ps -U myname | grep twm | nawk '{ print $1}'

store that pid number in a variable, then kill it....

i would appreciate any help....if i was unclear at all, let me know.

druuna · 10-06-2005, 10:37 AM

Hi,

VAR="`ps -U myname | grep twm | nawk '{ print $1}'`"
kill $VAR

Or, all in one go:

kill `ps -U myname | grep twm | nawk '{ print $1}'`

The backticks (upper left corner on most keyboards) around the command executes it and any output is given to kill (or a variable).

Hope this helps.

fatrandy13 · 10-06-2005, 10:42 AM

Thanks a lot. After messing around with your suggestion, and trying a few other things I came up with this:

ps -U myname | grep twm | cut -c1-6 | xargs kill > /dev/null

I appreciate the help.

montypython0102 · 09-13-2007, 07:59 PM

I tried this:

kill `ps -ef | grep processname | awk '{print $2}'`

The error that I keep getting is:

kill (12345): process does not exist.

This tells me that for some reason the pid gets enclosed in parens!

Any help?

druuna · 09-13-2007, 11:46 PM

Hi,

It is probably (also) trying to kill the grep process itself.

ps -ef | grep something will not only show the something process (if still active), but also the grep something part.

Try changing it to: kill `ps -ef | grep [Pp]rocessname | awk '{print $2}'`

Hope this helps.

BTW: Try starting a new thread next time. Although the problem seems simular, this could be seen as thread hijacking.

dive · 09-13-2007, 11:59 PM

You could also use 'pgrep' on the process name. This will not return the pid of grep.

kill `pgrep twm`