8 commands to effectively manage processes in Linux

Preface

The role of process management:

• Determine the health status of the server: Determine the server load and security status by analyzing the status of the process (memory, CPU occupancy, etc.)
• View all processes in the system
• Killing a process

This article will give you a detailed introduction to the Linux process management commands, and use these key commands to manage your applications throughout the process.

Generally speaking, there are three main states in the life cycle of an application process: starting, running, and stopping. Each state can and should be carefully managed if we want to be competent stewards. These eight commands can be used to manage the entire life cycle of a process.

Start the process

The easiest way to start a process is to type its name at the command line and press Return. If you want to start the Nginx web server, type nginx. Maybe you just want to see its version.

alan@workstation:~$ nginx

alan@workstation:~$ nginx -v
nginx version: nginx/1.14.0

Check your executable path

The above demonstration of launching a process assumes that the executable is in your executable path. Understanding this path is key to reliably launching and managing processes. Administrators will typically customize this path for their desired purposes. You can use echo $PATH to see your executable path.

alan@workstation:~$ echo $PATH
/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin

WHICH

Use the which command to view the full path of the executable file.

alan@workstation:~$ which nginx
/opt/nginx/bin/nginx

I will use the popular web server software Nginx as my examples. Assuming Nginx is installed. If the command which nginx returns nothing, then Nginx is not found because it only searches the executable path you specify. There are three ways to remedy the situation where a process cannot be started simply by name. The first is to type the full path -- although, I hate typing the whole thing, do you?

alan@workstation:~$ /home/alan/web/prod/nginx/sbin/nginx -v
nginx version: nginx/1.14.0

The second solution is to install the application in a directory that is in the executable's path. However, this may not always be possible, especially if you do not have root privileges.

A third solution is to update your executable path environment variable to include the installation directory of the specific application you want to use. This solution is shell dependent. For example, Bash users need to edit the PATH= line in their .bashrc file.

PATH="$HOME/web/prod/nginx/sbin:$PATH"

Now, repeat your echo and which commands or try to check the version. Much easier!

alan@workstation:~$ echo $PATH
/home/alan/web/prod/nginx/sbin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin

alan@workstation:~$ which nginx
/home/alan/web/prod/nginx/sbin/nginx

alan@workstation:~$ nginx -v            
nginx version: nginx/1.14.0

Keep the process running

NOHUP

When you log out or close the terminal, the process may not continue running. This special case can be solved by putting the nohup command in front of the command to be run to keep the process running. Additionally, appending an ampersand will send the process to the background and allow you to continue using the terminal. For example, suppose you want to run myprogram.sh.

nohup myprogram.sh &

nohup will return the PID of the running process. I'll talk more about PID next.

Managing running processes

Each process has a unique process identification number (PID). This number is what we use to manage each process. We can also use the process name, which I will demonstrate below. There are several commands to check the status of running processes. Let's take a quick look at these commands.

PS

The most common is the ps command. The default output of ps is a simple list of processes running in the current terminal. As shown below, the first column contains the PID.

alan@workstation:~$ ps
PID TTY TIME CMD
23989 pts/0 00:00:00 bash
24148 pts/0 00:00:00 ps

I want to see the Nginx process I started earlier. To do this, I tell ps to show me every running process ( -e ) and the complete listing ( -f ).

alan@workstation:~$ ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 Aug18 ? 00:00:10 /sbin/init splash
root 2 0 0 Aug18 ? 00:00:00 [kthreadd]
root 4 2 0 Aug18 ? 00:00:00 [kworker/0:0H]
root 6 2 0 Aug18 ? 00:00:00 [mm_percpu_wq]
root 7 2 0 Aug18 ? 00:00:00 [ksoftirqd/0]
root 8 2 0 Aug18 ? 00:00:20 [rcu_sched]
root 9 2 0 Aug18 ? 00:00:00 [rcu_bh]
root 10 2 0 Aug18 ? 00:00:00 [migration/0]
root 11 2 0 Aug18 ? 00:00:00 [watchdog/0]
root 12 2 0 Aug18 ? 00:00:00 [cpuhp/0]
root 13 2 0 Aug18 ? 00:00:00 [cpuhp/1]
root 14 2 0 Aug18 ? 00:00:00 [watchdog/1]
root 15 2 0 Aug18 ? 00:00:00 [migration/1]
root 16 2 0 Aug18 ? 00:00:00 [ksoftirqd/1]
alan 20506 20496 0 10:39 pts/0 00:00:00 bash
alan 20520 1454 0 10:39 ? 00:00:00 nginx: master process nginx
alan 20521 20520 0 10:39 ? 00:00:00 nginx: worker process
alan 20526 20506 0 10:39 pts/0 00:00:00 man ps
alan 20536 20526 0 10:39 pts/0 00:00:00 pager
alan 20564 20496 0 10:40 pts/1 00:00:00 bash

You can see the Nginx process in the output of the ps command above. This command displays nearly 300 lines, but I shortened it for this example. As you can imagine, trying to process 300 lines of process information is a bit confusing. We can pipe this output into grep to filter it to show only nginx.

alan@workstation:~$ ps -ef |grep nginx
alan 20520 1454 0 10:39 ? 00:00:00 nginx: master process nginx
alan 20521 20520 0 10:39 ? 00:00:00 nginx: worker process

It is indeed better. We can quickly see that Nginx has PIDs 20520 and 20521.

PGREP

The pgrep command further simplifies the problems encountered by calling grep alone.

alan@workstation:~$ pgrep nginx
20520
20521

Suppose you are in a hosting environment where multiple users are running several different Nginx instances. You can exclude others from the output using the -u option.

alan@workstation:~$ pgrep -u alan nginx
20520
20521

PIDOF

Another handy tool is pidof. This command will check the PID of a specific binary, even if another process with the same name is running. To set up an example, I copied my Nginx to a second directory and started it with the corresponding path prefix. In real life, this instance might be in a different location, such as a directory owned by a different user. If I run two Nginx instances, the pidof output shows all their processes.

alan@workstation:~$ ps -ef |grep nginx
alan 20881 1454 0 11:18 ? 00:00:00 nginx: master process ./nginx -p /home/alan/web/prod/nginxsec
alan 20882 20881 0 11:18 ? 00:00:00 nginx: worker process
alan 20895 1454 0 11:19 ? 00:00:00 nginx: master process nginx
alan 20896 20895 0 11:19 ? 00:00:00 nginx: worker process

Using grep or pgrep will display the PID numbers, but we may not be able to tell which instance is which.

alan@workstation:~$ pgrep nginx
20881
20882
20895
20896

The pidof command can be used to determine the PID of each specific Nginx instance.

alan@workstation:~$ pidof /home/alan/web/prod/nginxsec/sbin/nginx
20882 20881

alan@workstation:~$ pidof /home/alan/web/prod/nginx/sbin/nginx
20896 20895

TOP

The top command has been around for a long time and is very useful for viewing details of running processes and quickly identifying issues such as memory consumption. Its default view is shown below.

top - 11:56:28 up 1 day, 13:37, 1 user, load average: 0.09, 0.04, 0.03
Tasks: 292 total, 3 running, 225 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.1 us, 0.2 sy, 0.0 ni, 99.7 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 16387132 total, 10854648 free, 1859036 used, 3673448 buff/cache
KiB Swap: 0 total, 0 free, 0 used. 14176540 avail Mem

 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
17270 alan 20 0 3930764 247288 98992 R 0.7 1.5 5:58.22 gnome-shell
20496 alan 20 0 816144 45416 29844 S 0.5 0.3 0:22.16 gnome-terminal-
21110 alan 20 0 41940 3988 3188 R 0.1 0.0 0:00.17 top
 1 root 20 0 225564 9416 6768 S 0.0 0.1 0:10.72 systemd
 2 root 20 0 0 0 0 S 0.0 0.0 0:00.01 kthreadd
 4 root 0 -20 0 0 0 I 0.0 0.0 0:00.00 kworker/0:0H
 6 root 0 -20 0 0 0 I 0.0 0.0 0:00.00 mm_percpu_wq
 7 root 20 0 0 0 0 S 0.0 0.0 0:00.08 ksoftirqd/0

You can change the update interval by typing the letter s and the number of seconds you prefer for the updates. To make it easier to monitor our example Nginx process, we can call top with the -p option and pass the PID. This output is much cleaner.

alan@workstation:~$ top -p20881 -p20882 -p20895 -p20896

Tasks: 4 total, 0 running, 4 sleeping, 0 stopped, 0 zombie
%Cpu(s): 2.8 us, 1.3 sy, 0.0 ni, 95.9 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 16387132 total, 10856008 free, 1857648 used, 3673476 buff/cache
KiB Swap: 0 total, 0 free, 0 used. 14177928 avail Mem

 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
20881 alan 20 0 12016 348 0 S 0.0 0.0 0:00.00 nginx
20882 alan 20 0 12460 1644 932 S 0.0 0.0 0:00.00 nginx
20895 alan 20 0 12016 352 0 S 0.0 0.0 0:00.00 nginx
20896 alan 20 0 12460 1628 912 S 0.0 0.0 0:00.00 nginx

Correctly determining the PID is very important when managing processes, especially when terminating them. Furthermore, if top is used in this way, every time one of these processes stops or a new process starts, top needs to be informed of the new process.

Terminating a process

KILL

Interestingly, there is no stop command. In Linux, there is the kill command. kill is used to send a signal to a process. The most commonly used signals are "terminate" ( SIGTERM ) or "kill" ( SIGKILL ). However, there is more. Here are some examples. A complete list can be displayed using kill -L.

 1) SIGHUP 2) SIGINT 3) SIGQUIT 4) SIGILL 5) SIGTRAP
 6) SIGABRT 7) SIGBUS 8) SIGFPE 9) SIGKILL 10) SIGUSR1
11) SIGSEGV 12) SIGUSR2 13) SIGPIPE 14) SIGALRM 15) SIGTERM

Note that signal No. 9 is SIGKILL. Usually, we would issue a command such as kill -9 20896 . The default signal is 15, which is SIGTERM. Keep in mind that many applications have their own stopping methods. Nginx uses the -s option to pass signals, such as stop or reload. Generally, I prefer to use the application's specific method to stop an action. However, I will demonstrate using the kill command to stop the Nginx process 20896, and then use pgrep to confirm that it has stopped. PID 20896 no longer appears.

alan@workstation:~$ kill -9 20896
 
alan@workstation:~$ pgrep nginx
20881
20882
20895
22123

PKILL

The command pkill is similar to pgrep in that it can search by name. This means that you must be very careful when using pkill. In my Nginx example, if I only wanted to kill one Nginx instance, I probably wouldn't choose to use it. I can pass the Nginx option -s stop to the specific instance to eliminate it, or I need to use grep to filter the entire ps output.

/home/alan/web/prod/nginx/sbin/nginx -s stop
/home/alan/web/prod/nginxsec/sbin/nginx -s stop

If I want to use pkill, I can include the -f option to have pkill filter the entire command line arguments. This of course also applies to pgrep. So, before executing pkill -f , I can first confirm it with pgrep -a .

alan@workstation:~$ pgrep -a nginx
20881 nginx: master process ./nginx -p /home/alan/web/prod/nginxsec
20882 nginx: worker process
20895 nginx: master process nginx
20896 nginx: worker process

I can also use pgrep -f to narrow down my results. pkill with the same arguments will stop the process.

alan@workstation:~$ pgrep -f nginxsec
20881
           
alan@workstation:~$ pkill -f nginxsec

The key thing to remember about pgrep (and especially pkill) is that you must always ensure that your search results are accurate so that you don't accidentally affect the wrong process.

Most of these commands have many command line options, so I always recommend reading the man page for each command. Although most of these commands exist on platforms such as Linux, Solaris, and BSD, there are some differences. When working at the command line or writing scripts, always test and be prepared to make corrections as needed.

via: https://opensource.com/article/18/9/linux-commands-process-management

Author: Alan Formy-Duval Topic: lujun9972 Translator: heguangzhi Proofreader: wxy

Summarize

The above is the full content of this article. I hope that the content of this article will have certain reference learning value for your study or work. If you have any questions, you can leave a message to communicate. Thank you for your support for 123WORDPRESS.COM.