i'm root. i use ps to find out pid of processse i wanna kill. i type "kill xxx" but the process is still there ! what the heck ...!?

if you want to kill the program you dont need the pid in root. just type killall then the program name killall all together . if i want to stop my dialup program I type killall pppd

Try kill -9 'pid'. Kill uses TERM by default, and a running process might be hung and not respond to TERM, -9 is the KILL signal, and it terminates a process ungracefully.

thanks.

[root@localhost /]# ps
PID TTY TIME CMD
4662 pts/0 00:00:00 bash
5732 pts/0 00:00:00 yum
5764 pts/0 00:00:00 yum
6854 pts/0 00:00:00 man
6876 pts/0 00:00:00 sh
6878 pts/0 00:00:00 less
6900 pts/0 00:00:00 man
6921 pts/0 00:00:00 sh <defunct>
6923 pts/0 00:00:00 less
6978 pts/0 00:00:00 man
6999 pts/0 00:00:00 sh <defunct>
7001 pts/0 00:00:00 less
7305 pts/0 00:00:00 man
7326 pts/0 00:00:00 sh <defunct>
7520 pts/0 00:00:00 ps

i tried "kill -9 PID" and those i tried are shown "defunct" above. if i ps again, it still shows this screen. however, when i wrote C programs under unix days ago i only needed to use "kill PID" only. Also, once i killed a process, the screen didn't show the one just killed(i mean the system cleared that entry immediately w/o displaying <defunct>) if i entered ps again. why !?

thanks:

[root@localhost /]# ps
PID TTY TIME CMD
4662 pts/0 00:00:00 bash
5732 pts/0 00:00:00 yum
5764 pts/0 00:00:00 yum
6854 pts/0 00:00:00 man
6876 pts/0 00:00:00 sh
6878 pts/0 00:00:00 less
6900 pts/0 00:00:00 man
6921 pts/0 00:00:00 sh <defunct>
6923 pts/0 00:00:00 less
6978 pts/0 00:00:00 man
6999 pts/0 00:00:00 sh <defunct>
7001 pts/0 00:00:00 less
7305 pts/0 00:00:00 man
7326 pts/0 00:00:00 sh <defunct>
7780 pts/0 00:00:00 man
7801 pts/0 00:00:00 sh
7803 pts/0 00:00:00 less
7809 pts/0 00:00:00 man
7830 pts/0 00:00:00 sh
7831 pts/0 00:00:00 bzip2
7832 pts/0 00:00:00 less
7838 pts/0 00:00:00 man
7840 pts/0 00:00:00 sh
7842 pts/0 00:00:00 less
7852 pts/0 00:00:00 ps
[root@localhost /]# killall yum
[root@localhost /]# ps
PID TTY TIME CMD
4662 pts/0 00:00:00 bash
5732 pts/0 00:00:00 yum
5764 pts/0 00:00:00 yum
6854 pts/0 00:00:00 man
6876 pts/0 00:00:00 sh
6878 pts/0 00:00:00 less
6900 pts/0 00:00:00 man
6921 pts/0 00:00:00 sh <defunct>
6923 pts/0 00:00:00 less
6978 pts/0 00:00:00 man
6999 pts/0 00:00:00 sh <defunct>
7001 pts/0 00:00:00 less
7305 pts/0 00:00:00 man
7326 pts/0 00:00:00 sh <defunct>
7780 pts/0 00:00:00 man
7801 pts/0 00:00:00 sh
7803 pts/0 00:00:00 less
7809 pts/0 00:00:00 man
7830 pts/0 00:00:00 sh
7831 pts/0 00:00:00 bzip2
7832 pts/0 00:00:00 less
7838 pts/0 00:00:00 man
7840 pts/0 00:00:00 sh
7842 pts/0 00:00:00 less
7856 pts/0 00:00:00 ps
[root@localhost /]#

well.... yum's are still there. any ideas !?

You shouldn't have to remember 'kill -9', it's the same as 'kill -KILL' and the pid or you can use 'killall -KILL' and program name. If you wanted to kill all instances of yum you would run:

as root.

thanks. it worked, too. i think both kill -9 xxx and killall -KILL name are feasible. however, i still wonder why i just needed to type kill pid under unix w/0 including -9 ???

As mentioned further up, normally
kill <pid>
is sufficient to tell a prog to close GRACEFULLY if you own it or you are root.
For emergencies and or if the prog is 'hung',
kill -9 <pid>
should kill it UNGRACEFULLY ie you risk open/corrupted files, broken sockets etc.
Occasionally you get processes so badly hung that they remain zombies until a reboot. This is rare, but has been known.