How processes are managed in Linux
A running program is called a process. If we talk about the terminal, the processes can either be running interactively in the foreground or they can be sent in the background to execute. A process which is sent to the background is referred to as a job. We can send processes in the background during runtime, we can bring background processes to the front and we can send various types of signals to the running processes to change their execution states.
The complete list of signals which can be sent to a process are given in the table below:
The complete list of signals which can be sent to a process are given in the table below:
Signal Name
Number
Description
SIGHUP
1
Hangup (POSIX)
SIGINT
2
Terminal interrupt (ANSI)
SIGQUIT
3
Terminal quit (POSIX)
SIGILL
4
Illegal instruction (ANSI)
SIGTRAP
5
Trace trap (POSIX)
SIGIOT
6
IOT Trap (4.2 BSD)
SIGBUS
7
BUS error (4.2 BSD)
SIGFPE
8
Floating point exception (ANSI)
SIGKILL
9
Kill(can’t be caught or ignored) (POSIX)
SIGUSR1
10
User defined signal 1 (POSIX)
SIGSEGV
11
Invalid memory segment access (ANSI)
SIGUSR2
12
User defined signal 2 (POSIX)
SIGPIPE
13
Write on a pipe with no reader, Broken pipe (POSIX)
SIGALRM
14
Alarm clock (POSIX)
SIGTERM
15
Termination (ANSI)
SIGSTKFLT
16
Stack fault
SIGCHLD
17
Child process has stopped or exited, changed (POSIX)
SIGCONT
18
Continue executing, if stopped (POSIX)
SIGSTOP
19
Stop executing(can’t be caught or ignored) (POSIX)
SIGTSTP
20
Terminal stop signal (POSIX)
SIGTTIN
21
Background process trying to read, from TTY (POSIX)
SIGTTOU
22
Background process trying to write, to TTY (POSIX)
SIGURG
23
Urgent condition on socket (4.2 BSD)
SIGXCPU
24
CPU limit exceeded (4.2 BSD)
SIGXFSZ
25
File size limit exceeded (4.2 BSD)
SIGVTALRM
26
Virtual alarm clock (4.2 BSD)
SIGPROF
27
Profiling alarm clock (4.2 BSD)
SIGWINCH
28
Window size change (4.3 BSD, Sun)
SIGIO
29
I/O now possible (4.2 BSD)
SIGPWR
30
Power failure restart (System V)
Foreground and Background Process Manipulation
Execute a process with the & sign appended to the command-line at the end will send the process to the background.
sleep 1000 &
sleep 2000 &
If we have a running process in the foreground, follow these steps to send it to the background.
sleep 3000
Press ctrl+z
The process is now suspended. Type:
jobs
This will list the active jobs in the terminal. Find the number of the suspended job (assumed to be 3 for our example) and type:
bg %3
To bring the same job to the foreground, type:
fg %2
Just entering fg will bring the most recent job to the foreground.
Disassociating Processes from the Terminal
A problem arises when we close the terminal. All background jobs are closed as closing sends all tasks a SIGHUP signal. We can check the running processes by issuing the following command, first before closing the terminal and afterwards when we open a new terminal.
ps -aux | grep sleep
As can be seen, all jobs are now closed due to the terminal closure. So what can we do to protect the processes from automatically closing?
Protecting tasks from SIGHUP
We can execute a process using the nohup program. This will allow the program to ignore SIGHUP when the terminal is closed.
nohup sleep 1000 &
If we do not use nohup and we want to achieve the same after running a process, we can use the disown command on a job number.
disown %3
Nohup has a problem as it redirects the output to nohup.out. This means that several processes started from the same directory will have a common nohup file making matters complex. The other problem is that a job created using nohup or disown’ed after execution cannot be brought to the foreground.
nohup sleep 9999 &
Close terminal and open
ps aux | grep sleep
sleep 8888 &
disown %1
When we disconnect, we cannot ‘fg’ it. This is a problem if we want to interact with the process. To overcome this we will now use a nifty utility called screen.
Screen
Install the screen program, if it is not already installed by typing:
apt-get install screen
Or
yum install screen
Open two terminals and in one of the terminals type:
screen
Then run a sleep program in this terminal on the foreground by typing:
sleep 99999
Press
Ctrl + a and then press the ‘d’ key to disconnect the screen session.
ps aux | grep sleep
will show you that sleep is still running.
In the other terminal, type
screen –dr
Will give us back the session on this other terminal. Close all terminals. Go to any other terminal and type:
screen –dr
This is just a tip of the iceberg of what screen can do. Please refer to screen documents to see the awesome power which screen provides to the system administrator.
Process IDs and Signals and the kill Command
When a process is created, it is assigned a PID which is just an identifier for each process running on the system. We can send signals to the process using this PID. Common signals were listed in the table in the start of this section.
The format of sending a signal is to issue the kill command with the necessary signal.
kill –SIGNAL_NUMBER PID
The default behavior of the kill command is to send the SIGTERM signal which lets the process to gracefully exit. This is the proper way to kill the process. If the process doesn’t honor the SIGTERM signal due to a locking problem, send SIGKILL which will almost always end a process. When we press Ctrl+C, a SIGINT is sent to the process. SIGHUP is more like a courtesy to inform the process that its parent is closing, so it should close itself.
The following commands can be used to test various kill commands.
Sleep 1000
Ctrl + C
sleep 1000 &
kill PID
Default kill signal is SIGTERM
sleep 1000 &
kill -9 PID
ps –aux | grep sleep
Advanced process management
We will now see some additional commands for process management namely killall, pgrep and pkill. The killall command will kill all processes with the same name. It requires that an exact match is provided for the command to work.
sleep 1000 &
sleep 1000 &
ps –eaf | grep sleep
killall sleep
killall -9 sleep
The killall command will not be able to kill other users’ processes unless we are root.
The pkill and pgrep provide a safety net against killing arbitrary processes. They don’t require the whole process name, rather a part of the process name will suffice for them to function. Additionally pgrep will catch child process with the same name. Other commands only work on the parent processes and don’t list any children with the same name.
pgrep sle
pgrep –a sle
Make a bash script sleepytime
vi sleepytime
#!/bin/bash
sleep 12345
Now the sleep process is not listed in the pkill command, so we will use pgrep.
ps –eaf | grep sleep
pgrep –a slee
pgrep –af slee
Process Priorities and Niceness
Each process on a Linux is assigned a priority. The priority can be thought of as the portion of CPU time in an interval which a process is entitled to. The higher the priority, the larger the time slice and vice versa. In Linux the priority is called the niceness level of a process. The nice levels are from -20 to 19, where -20 is the highest priority and 19 is the lowest. Each process is created with a default priority of 0. The command to manipulate nice levels are:
nice
renice
Starting Levels
To start a new process with a custom priority level use:
nice -10 application
nice — 15 application
nice -19 application
nice — 20 application
Please note that if a lower priority is to be assigned, then a single dash (-) followed by a number is given as the argument. If a priority less than 0 is to be assigned then a double dash is typed ( — ). Only the root user can increase priority of a process.
sleep 1111 &
nice -15 sleep 22222 &
nice — 10 sleep 1111 &
ps –alf
If we are not allowed to set the nice level, the command is still executed with the default niceness level. Now kill all sleep processes using:
pkill –f slee
Changing Nice Levels
Once a process is started, its niceness level can be changed using the renice command. Remember that only root user can increase the priority level of a process. If priority is lowered, we cannot mark it higher to even 0, which is the default.
sleep 1111 &
renice 19 -p PID
renice -10 -p PID; #Won’t work unless root
renice 15 -p PID; #Won’t work unless root and the priority was changed to 19 first
