Basically I want to dynamically start some processes which may create their own children processes, also I want to kill a certain group of processes I just created whenever I want. One way I could think of is to start processes with a name(to distinguish as a group), then use pkill to kill them by the name. The question is how to start a process with a name so that I can use pkill to kill them by the name? I am open to other solutions as well.
How can I start a process with a different name?
bash -c "exec -a <MyProcessName> <Command>"
Then you can kill the process with:
pkill -f MyProcessName
Related
When I want to kill a process using the pid in linux, its subprocess still existes. I hope to kill all process using one command.
Suggesting command pkill -p PID pattern .
See more documentation here.
Check out process groups:
https://en.wikipedia.org/wiki/Process_group
Assuming you want to do this from a shell?
If you do a kill and make the top process negative it does a killpg under the covers and sends the signal to all the processes in the group.
I have a batch file in linux (which I will execute externally from within my lazarus application). What it should do is read a process PID, store it in a variable, and use that variable to execute the "kill" command.
This is how I'm doing it:
PID=`pidof myProcess`
kill $PID
However, the kill command fails with a ": arguments must be process or job IDs" error.
How can I achieve this?
Perhaps using pkill directly would better suit your needs.
pkill myProcess
More info on pkill here: https://www.lifewire.com/list-and-kill-processes-using-the-pgrep-and-pkill-4065112
I have seen many discussions here on kill command. But my confusion is different. I have many processes with the same name and I have to automate the killing. Hence I can't use the pid. So is it possible that if I go to a specific path and use kill <pname> then only the process related to that path will get killed?
Or is there some way to incorporate the path name in kill command?
Instead of using a pid, you could always use the pkill command and have it check against some regular expression. If you pass it the -f flag, it allows you to check against the entire command line rather than just the process name.
Something like this would probably do the trick:
pkill -TERM -u username -f "mwhome.*weblogic\\.NodeManager"
-f is where you would pass in your regex
-u is also useful so that you only affect pid's running as specific users
No, but you when you start the process with
yourcommand & echo $!
or wrap it in a small script
#!/bin/bash
yourcommand &
echo $! >/path/to/pid.file
you can save the pid. And then kill the process with this pid. This is the normal way how to manage the processes. If you look in the normal init.d scripts of perhaps nginx, they do it the same way. Just saving the pid in a file, and at stopping just read the pid and kill the process.
I al running this command :
pgrep -l someprocess
I get some outputs XXXX someprocess
then I kill every process appearing by hand, I would like to write a script to do it automatically, but this doesn(t make sense
kill -9 $(pgrep -l someprocess | grep "^[0-9]{4}")
someone could help ?
You can use either pkill or killall to accomplish exactly that.
I found this short and clear summary explaining the different ways of killing processes.
pkill is as simple as: pkill someprocess.
#ewm already included a detailed explanation about killall in his answer, so I'm not repeating it here.
You might want to look at the 'killall' command:
KILLALL(1) User Commands KILLALL(1)
NAME
killall - kill processes by name
SYNOPSIS
killall [-Z,--context pattern] [-e,--exact] [-g,--process-group] [-i,--interactive] [-q,--quiet]
[-r,--regexp] [-s,--signal signal] [-u,--user user] [-v,--verbose] [-w,--wait] [-I,--ignore-case]
[-V,--version] [--] name ...
killall -l
killall -V,--version
DESCRIPTION
killall sends a signal to all processes running any of the specified commands. If no signal name is
specified, SIGTERM is sent.
Signals can be specified either by name (e.g. -HUP or -SIGHUP ) or by number (e.g. -1) or by option
-s.
If the command name is not regular expression (option -r) and contains a slash (/), processes execut-
ing that particular file will be selected for killing, independent of their name.
killall returns a zero return code if at least one process has been killed for each listed command,
or no commands were listed and at least one process matched the -u and -Z search criteria. killall
returns non-zero otherwise.
A killall process never kills itself (but may kill other killall processes).
I have a script that runs from a newly created shell.
OS is Red Hat Enterprise Linux Server release 5.4 (Tikanga).
In certain point when the script detects that some application (started by the script) is hanging the script tries to terminate all the procesess it started. I assumed that the correct command for terminating all processes started by current user in current shell is:
pkill /?
The problem is that it kills sshd that is started in parent shell (by init.d) and the putty console disconnects showing error message.
I wonder:
How is it possible for specific user in specific shell to terminate process started by other user in parent shell?
What would be correct command to terminate all processes started by the script currently running?
I have found some solution where i store all the PIDs and when the script needs to terminate them I run in the loop the following:
[ws#RHDev ~]# pkill $(ps aux | grep $pid | awk '{print $2}')
However, I am looking for one-liner that simply terminates all the processes started by the current script.
You can filter subprocesses by the current process pid. The ps command can do this using the --ppid parameter.
ps opid --ppid=7051 | tail -n +2 | xargs kill
here tail -n +2 is to strip the ps headers.
I assumed that the correct command for terminating all processes started by current user in current shell is:
pkill /?
This command matches every single process in the system because it effectively requires only 0 symbols from process name to match. pgrep -l /? demonstrates that.
How is it possible for specific user in specific shell to terminate process started by other user in parent shell?
From man kill(2):
For a process to have permission to send a signal it must either be
privileged (under Linux: have the CAP_KILL capability), or the real or
effective user ID of the sending process must equal the real or saved
set-user-ID of the target process. In the case of SIGCONT it suffices
when the sending and receiving processes belong to the same session.
Do you invoke pkill from user root?
What would be correct command to terminate all processes started by the script currently running?
When bash starts it creates its own process group. Child processes it creates are put in the same process group. Hence:
kill -- -$$
Kills all processes started by the current script. Provided that those processes didn't become group leaders.