Commands

Commands

&

The & command, used on the command line, forces the specified process to run in the background.

Example: To start a lengthy program running in the background, use

sort_everything &

As soon as they are successfully invoked, background processes release control of your terminal to the shell that invoked them. This means that you can continue to work without waiting for a lengthy process to finish.

 

^Z

The ^Z command stops the current foreground process.

Example: This command works on any process. For example, say you're running a lengthy sort and you want to check what time it is:

$> biglongsort	# Start lengthy process
^Z	# Suspend with ^Z
[3] Stopped (signal)	# Shell acknowledges stop signal ^Z
$> date	# Run some other command
$> fg 3	# Restart the process

bg

job_num

The bg command takes a stopped job and starts it in the background.

Example: Say you started running a find command and it was taking forever. You've got only one window available, so you want to move it to the background without restarting it:

$ find / -type f -print | grep sammich
^Z	# Stop current job
$ jobs	# Display job list
[2]+ Stopped find / -type f -print | grep sammich
$ bg 2	# Run job in background

 

To use this command, you will need to know the job number. Obtain the job number with the jobs command.

 

fg

job_num

The fg command moves a running background job to the foreground.

Example: To get the job numbers of your background and stopped jobs, start bigsort running in the foreground:

$ jobs	# Tell system to display job list
[2] Running bigsort
[3] Stopped find / -type f -print | grep sammich
$ fg 2	# Move bigsort to foreground

 

To use this command, you must know the job number. Obtain the job number with the jobs command.

 

fuser

[-a|-s] [-n space] [-signal] [-kmuv] name [-][-n space] [-signal] [-kmuv] name

fuser -l

fuser -V

The fuser command is used to display the PID of any process(es) using a specified file. The various types of access are indicated as follows:

c	File is process' current directory.
e	Process is running file as executable.
f	File is open for modification.
r	File is process' root directory.
m	Mmap'ed file or shared library.

Example: To find and kill all processes running the file /bin/program, use

fuser -k /bin/program

-a	Show all files listed on the command line, whether accessed or not.
-k	Attempt to kill any process accessing the file.
-l	List signal names.
-m	If the file listed is s mounted filesystem, lists all processes accessing members of that filesystem.
-n	Space select a different name space.
-s	Silent operation.
-signal	Use the specified signal instead of SIGKILL when killing processes.
-u	Include the user name of the process' owner in output.
-v	Verbose output. Include PID, USER, COMMAND.
-V	Output version number.
-	Return all options to their default values.

jobs

The jobs command displays a list of the current background processes.

Example: Say you had two jobs bigsort and a find running in the background. If so, jobs might display the following output:

$ jobs
[2] Running	bigsort
[3] Stopped	find / -type f -print | grep sammich

kill

[ -s signal | -p ] [ -a ] pid

kill -l

[ signal ]

Kill a process. More specifically, the kill command sends one of a variety of possible signals to the specified process. The default signal is TERM. The pid parameter is a process id, which may be obtained using the ps command. You need to be superuser to use this command on anyone other than yourself.

Example: To kill process 1234 immediately without giving it a chance to finish what it is doing, use

kill -9 1234

One way to kill processes is by process id number. To get the process id number, run ps aux and grep for the name of the command you're targeting. Another is to kill by job number:

 

kill %1

To kill someone else's process, you must be a superuser. The command to use is

 

-s signal

Use the specified signal, rather than the default.

-p	Print the pid of the named process, but do not signal it.
-l	Output a list of possible signals.

Example: To display the list of all possible signals, use the following (see Figure 3-1 for a tidy, formatted version of the output of the following command):

kill -l

Figure 3-1. Signals

killall

[-eiqvw] [-signal] command_name

killall

-l

killall

-V

killall attempts to kill all processes executing the command specified by command_name. By default, this is accomplished by use of the SIGTERM signal. However, other signals may be specified either by name (see /usr/include/linux/signal.h) or by number.

Example: To kill all processes (not just those owned by you) currently running the mail command, use

killallmail

Example: To kill all processes running the mail command with the strongest possible signal, use

killall -9 mail

-e	Requires long filenames ( > 15 char) to be matched exactly.
-i	Prompts the user for confirmation before killing.
-l	Lists possible signals.
-q	Runs quietly. Suppress error messages.
-v	Runs verbosely. Display a message for each successful termination.
-V	Outputs version information.
-w	Waits for signaled processes to die.

nice

[OPTION] [COMMAND [ARG] ]

Start a command with the specified priority rather than the default. By default, commands are run with a priority of zero. To give your command a lower priority (and thereby make it run slower), you may specify a value between 1 and 19. To make it run faster, you may specify a value between -1 and -20. You must be the superuser to specify a negative value.

Example: To start up big_program but force it to yield some of its processor time to other programs, use

nice 19 big_program

Example: To start up big_program, but enable it to snatch some processor time from other, less deserving programs, use

nice -20 big_program

Use renice to modify the priority of a command which is already running.

 

To start a process with a higher priority (e.g., a lower priority number), you must be a superuser.

 

-nADJUSTMENT
-ADJUSTMENT --adjustment=ADJUSTMENT	Use the specified ADJUSTMENT instead of the default.

nohup

COMMAND [ARG]

Execute the specified command, ignoring the hangup signals (particularly those sent when you logout). The idea is that the command will continue to run after you logout. Also, nohup increases the scheduling priority to 5, which means it will run a bit slower. Output is sent to nohup.out.

Example: To start up big_program telling it to continue to run even after you're at home asleep, use

nohup big_program

pidof

[-s] [-x] [-o omitpid] [-o omitpid..] program [program..]

Locate and print the process id of the specified program.

Example: To learn the process id of the sendmail program, use

pidof sendmail

-s	Limit output to a single pid.
-x	If "program" is a script, also output the pid's of shells running the script.
-o omitpid	Ignore processes with the specified pid.

ps

[-] [lujsvmaxScewhrnu] [txx] [O[+|-]k1[[+|-]k2 ]] [pids] --sortX[+|-]key[,[+|-]key[, ]], --help, --version

ps (Process Status) retrieves and displays information about currently running processes. Depending on the command line options, the output of this command will include various permutations of the following:

Column	Description
%CPU	Rough percentage of CPU time consumed.
%MEM	Rough percentage of system memory consumed.
COMMAND	Command that started process.
NI	Nice value of process.
PID	Process ID.
PPID	Process Parent ID.
RSS	Real memory used.
SIZE	Virtual memory consumed by process.
STA	Current process state. One of the following:
R	Running
S	Sleeping
D	Uninterruptible sleep
T	Stopped or Traced
Z	Zombie
START	Time process started.
TIME	Total CPU time used.
TTY	Terminal associated with process.
UID	User id of process originator.
USER	User name of process originator.
WCHAN	Event process is waiting for.

Example: To display a long list of information about my current processes, use

ps l

FLAGS  UID  PID  PPID PRI  NI  SIZE  RSS WCHAN  STA TTY TIME COMMAND

  100  0  390  1  0  0  1232  808 wait4  S  1  0:00 -bash

100100  0  391  1  0  0  724  296 read_chan  S  2  0:00 /sbin/min

100100  0  392  1  0  0  724  296 read_chan  S  3  0:00 /sbin/min

100100  0  393  1  0  0  724  296 read_chan  S  4  0:00 /sbin/min

100100  0  394  1  0  0  724  296 read_chan  S  5  0:00 /sbin/min

100100  0  395  1  0  0  724  296 read_chan  S  6  0:00 /sbin/min

100000   0  420  390  0  0  1200  648 wait4  S  1  0:00 sh /usr/X

  100  0  421  420  0  0  1828  656 wait4  S  1  0:00 xinit /us

100100  0  424  421  0  0  2968  1944 do_select  S  1  0:01 /usr/X11R

100100  0  433  424  0  0  1496  724 nanosleep  S  1  0:00 wmclock -

100100  0  878  424  0  0  2488  1656 do_select  S  1  0:00 xterm -sb

100100  0  970  424  0  0  2448  1620 do_select  S  1  0:00 xterm -sb

  100  0  879  878  0  0  1240  808 wait4  S  p1  0:00 bash

100100  0  884  879  0  0  1244  768 do_signal  T  p1  0:00 vi

100100  0  911  879  0  0  1268  720 do_signal  T  p1  0:00 top

100100  0  1016  879  0  0  1388  752 do_select  S  p1  0:00 vi t

  100  0  971  970  10  0  1236  796 wait4  S  p2  0:00 bash

100000  0  1020  971  11  0  956  492  R  p2  0:00 ps l

Example: To display full process information about all processes, use

ps aux

Example: To display full (f option) process information about all processes associated with user jones (u option) to standard output, use

ps fu jones

l	Long format.
u	User format include user name and start time.
j	Jobs format pgid sid.
s	Signal format.
v	Vm format.
m	Displays memory information (combine with p flag to get number of pages).
f	"Forest" family tree format for command line.
a	Show processes of other users, too.
x	Show processes without controlling terminal.
S	Add child cpu time and page faults.
c	Command name from task_struct.
e	Show environment after command line and "+".
w	Wide output: Don't truncate command lines to fit on one line.
h	No header.
r	Running processes only.
n	Numeric output for USER and WCHAN.
txx	Look only at processes with controlling tty xx.
O[+|-]k1[,[+|-]k2[,...]]	Order the output by the sequence of sort keys k1, k2, - reverses default order of sort. There are a lot of possible sort keys.

 

Short	Long	Description
c	cmd	Simple name of executable.
C	cmdline	Full command line.
f	flags	Flags as in long format F field.
g	pgrp	Process group ID.
G	tpgid	Controlling tty process group ID.
j	cutime	Cumulative user time.
J	cstime	Cumulative system time.
k	utime	User time.
K	stime	System time.
m	min_flt	Number of minor page faults.
M	maj_flt	Number of major page faults.
n	cmin_flt	Cumulative minor page faults.
N	cmaj_flt	Cumulative major page faults.
o	session	Session ID.
p	pid	Process ID.
P	ppid	Parent process ID.
r	rss	Resident set size.
R	resident	Resident pages.
s	size	Memory size in kilobytes.
S	share	Amount of shared pages.
t	tty	The minor device number of tty.
T	start_time	Time process was started.
U	uid	User ID number.
u	user	User name.
v	vsize	Total VM size in bytes.

 

Sort order
-Ou , --sort:user	-Oc , --sort:cmd
-OU , --sort:uid	-Op , --sort:pid
-OP , --sort:ppid	-Og , --sort:pgrp
-Oo , --sort:session	-Ot , --sort:tty
-OG , --sort:tpgid	-Ok , --sort:utime
-OK , --sort:stime	-Oj , --sort:cutime
-OJ , --sort:cstime	-Oy , --sort:priority
-OT , --sort:start_time	-Of , --sort:flags
-Om , --sort:min_flt	-On , --sort:cmin_flt
-OM , --sort:maj_flt	-ON , --sort:cmaj_flt
-Ov , --sort:vsize	-Or , --sort:rss
-OC , --sort:pcpu	-Os , --sort:size
-OR , --sort:resident	-OS , --sort:share

 

xpids	Restrict output to only the listed processes.
--help	Output a help message.
--version	Display version information.

pstree

[-a] [-c] [-h] [-l] [-n] [-p] [-u] [-G|-U] [pid|user]

pstree

-V

Output the family tree of running processes. If no process id or user name is specified, the tree is rooted at init. Identical branches are merged in the display by prefixing the process name with a count.

Example: Say your xinit process has a PID of 430. To display it and all its descendants in tree form with their PIDs, use

pstree 430 -p

xinit(431)-+-X(432)

  `-wmaker(434)-+-wmclock(443)

  |-xterm(637)---bash(638)

  |-xterm(715)---bash(716)---pstree(2118)

  |-xterm(718)---bash(719)---top(720)

  |-xterm(721)---bash(722)---vi(2051)

   `-xterm(724)---bash(725)

-a	Include command line arguments in the display.
-c	Do not compact the identical subtrees.
-G	Graphical display. (Use vt100 line drawing characters.) Very cool.
-h	Highlight family tree of current process.
-l	Do not truncate long lines.
-n	Numeric sort by PID.
-p	Include PIDs in display.
-u	Include uid transitions in display.
-U	Graphical display. Uses UTF-8 (Unicode) graphical characters in display.
-V	Display version information.

renice

priority [[-p] pid ] [[-g] pgrp ] [[-u] user ]

renice enables you to alter the scheduling priority of running process(es). The default priority is 0. To make a process run faster, specify a negative priority value (you must be a superuser to do this). To give a process a lower priority (and thereby make it run slower), specify a positive priority. The priority value must be in the range -20 <= priority <= 19.

Example: To play a funny joke on your buddy, "Tim," and make all his stuff run really slowly, use

renice 20 -u tim

You must be a superuser to be this funny.

To give a running process a higher priority (by decreasing its priority number) you must be a superuser.

 

-g	Include processes belonging to the specified group in any modifications.
-u	Include processes belonging to the specified user in any modifications.
-p	Include the specified process in any modifications.

top

[-] [d delay] [q] [c] [S] [s] [i]

Top displays a listing of the processes utilizing the CPU. Output may be ordered by CPU usage, memory usage or runtime. Note that top is itself a CPU-intensive task, so use it sparingly. To change options while top is running, use the s command at the prompt and specify the new options.

Example: In my experience, top is generally run without arguments.

top

It produces output such as the following:

8:59am  up 3 days, 16:42,  3 users,  load average: 0.00, 0.00, 0.00

39 processes: 38 sleeping, 1 running, 0 zombie, 0 stopped

CPU states:  0.9% user,  0.0% system,  0.0% nice, 99.1% idle

Mem:  95460K av,  87380K used,  8080K free,  17648K shrd,  33748K buff

Swap: 104384K av,  0K used, 104384K free  37184K cached

  PID USER  PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM  TIME COMMAND

 6543 root  19  0  716  716  556 R  0  0.9  0.7  0:00 top

  1 root  0  0  388  388  328 S  0  0.0  0.4   0:03 init

  2 root  0  0  0  0  0 SW  0  0.0  0.0  0:00 kflushd

  3 root  -12 -12  0  0  0 SW<  0  0.0  0.0  0:00 kswapd

  4 root  0  0  0  0  0 SW  0  0.0  0.0  0:00 md_thread

  5 root  0  0  0  0  0 SW  0  0.0  0.0  0:00 md_thread

  402 news  19  0  708  708  576 S  0  0.0  0.7  0:00 innwatch

 1287 root  0  0  796  796  656 S  0  0.0  0.8  0:00 bash

  395 root  0  0  296  296  248 S  0  0.0  0.3  0:00 mingetty

  36 root  0  0  364  364  312 S  0  0.0  0.3  0:00 kerneld

  210 bin  0  0  332  332  264 S  0  0.0  0.3  0:00 portmap

  224 root  0  0  472  472  396 S  0  0.0  0.4  0:00 syslogd

  233 root  0  0  536  536  324 S  0  0.0  0.5  0:00 klogd

  244 daemon  0  0  404  404  328 S  0  0.0  0.4  0:00 atd

  255 root  0  0  472  472  392 S  0  0.0  0.4  0:00 crond

  267 root  0  0  396  396  328 S   0  0.0  0.4  0:00 inetd

  278 root  0  0  1168 1168  580 S  0  0.0  1.2  0:00 named

-d	Use the specified delay (in seconds) between updates.
-q	Refresh constantly with no delay.
-S	Include in usage calculations the usage of the process' children.
-s	Run in secure mode.
-i	Tells top to ignore idle or zombie processes.
-c	Include the process' entire command line in the output.

Summary of Display Values

Uptime	How long the system has been running, plus the average number of processes ready to run in the last 1, 5, and 15 minutes.
Processes	Count of the number of processes running at the last update.
CPU states	Percentage of CPU time in user mode, system mode, niced tasks and idle.
Mem	Memory statistics (total K, free K, used K, shared K, buffer K).
Swap	Swap statistics (total K, available K, used K, cached).
PID	PID of the process.
PPID	PID of the process' parent.
UID	UID of the process' owner.
USER	User name of the process' owner.
PRI	Priority of the task.
NI	Nice value of the task.
SIZE	Size of code + data + stack space, in kilobytes.
TSIZE	Code size of the task.
DSIZE	Size of data + stack space.
TRS	Text Resident Size.
SWAP	Size of part of task currently swapped out.
D	Size of dirty memory pages.
LIB	Size of library pages.
RSS	Total amount of physical memory used by the task.
SHARE	Amount of shared memory used by the task.
STAT	State of the task (Z=zombie, S=sleeping, T=stopped).
WCHAN	Show the address or name of the kernel function task is sleeping in.
TIME	Total CPU time used by task since invoked.
%CPU	Percentage of CPU time used by task since last update.
%MEM	Percentage of physical memory used by task.
COMMAND	Name of command which started task.

Once invoked, top may be used interactively. The following are supported:

<space>	Update display information immediately.
^L	Repaint the screen.
h, ?	Display help information.
k	Kill a process: Top will prompt for pid.
i	Ignore zombie processes and idle processes.
n, #	Show the specified number of processes.
q	Quit.
r	Renice a process.
S	Toggle cumulative mode on/off.
s	Specify new delay (in seconds) between updates.
f, F	Add specified fields to display.
o, O	Change order of displayed fields.
l	Toggle display of load average and uptime information.
m	Toggle display of memory information.
t	Toggle display of CPU state and process information.
c	Toggle display between command name/full command line.
M	Order display by memory usage.
P	Order display by CPU usage.
T	Order display by time/cumulative time.
W	Write current setup to ~/.toprc.