Shell Programming and Scripting

Hi
I need to write a small script to kill the process id of particular job in one shot ,
Example

Code:

> ps
 PID TTY       TIME COMMAND
 16280 pts/70    0:00 sh
 16278 pts/70    0:00 rlogind
 16197 pts/70    0:00 ps
 1234  pts/70    0:00  runflow
 2341  pts/70    0:00  runflow
 12673 pts/70    0:00  runflow.

I need to kill runflow process id .Everytime to write all the prcosess id's of runflow is quite tedious as sometimes there are lot of runflow scripts running .

Code:

kill 1234 2341 12673

i tried to do but i am stuck ..

Code:

ps |grep "runflow" | cut -f2 -d' '

This will show the process id of run jobs , but i need all this to be arranged in a row and then i need to kill the process.
Please suggest .

Moderator's Comments:

Use code tags please.

Hi
Are you looking for something like this:

Code:

$ ps | awk '/runflow/{print "kill -9 ", $1}'
kill -9  1234
kill -9  2341
kill -9  12673

This output can be copied to a file, and you can run the file or you can run the command itself in backticks.

Guru.

Code:

ps -ef| awk '$NF ~ /runflow/ {print $2}'| xargs kill

Can also use the system() function inside awk to issue the kill. xargs would be redundant then.

Code:

pkill runflow

Why not just use "killall runflow"?

edit: or pgrep

Code:

pgrep runflow | xargs kill

IMO the awk solutions just like greps need to use /[r]unflow/ to prevent the awk process itself from being killed.

---------- Post updated at 08:10 ---------- Previous update was at 08:06 ----------

Not all systems have killall and/or pgrep/pkill. Be careful where you run the killall command as it has a different meaning (oops) on at least an HPUX system:
killall(1M)

On AIX for example killall is not the same as on Linux. That's why it could be a reason to use it not for this purpose.