Introduction to Unix commands

This is a very brief introduction to some useful Unix commands, including examples of how to use each
command. For more extensive information about any of these commands, use the man command as described
below. Sources for more information appear at the end of this document.
Commands

cal cat cd chmod

cp date df du

find jobs kill less and more

lpr and lp ls man mkdir

mv ps pwd rm

rmdir set vi w and who

cal
This command will print a calendar for a specified month and/or year.
To show this month's calendar, enter:
cal
To show a twelve-month calendar for 2004, enter:
cal 2004
To show a calendar for just the month of June, 1970, enter:
cal 6 1970
For more detailed information, see the Knowledge Base document In Unix, how can I display a calendar?

cat
This command outputs the contents of a text file. You can use it to read brief files or to concatenate files
together.
To append file1 onto the end of file2, enter:
cat file1 >> file2
To view the contents of a file named myfile, enter:
cat myfile
Because cat displays text without pausing, its output may quickly scroll off your screen. Use the less command
(described below) or an editor for reading longer text files.
For more detailed information, see the Knowledge Base document In Unix, how do I combine several files into
a single file?

cd
This command changes your current directory location. By default, your UNIX login session begins in your
home directory.
To switch to a subdirectory (of the current directory) named myfiles, enter:
cd myfiles
To switch to a directory named /home/dvader/empire_docs, enter:
cd /home/dvader/empire_docs
To move to the parent directory of the current directory, enter:
cd ..
To move to the root directory, enter:
cd /
To return to your home directory, enter:
 cd
chmod
This command changes the permission information associated with a file. Every file (including directories,
which Unix treats as files) on a Unix system is stored with records indicating who has permission to read,
write, or execute the file, abbreviated as r, w, and x. These permissions are broken down for three categories
of user: first, the owner of the file; second, a group with which both the user and the file may be associated;
and third, all other users. These categories are abbreviated as u for owner (or user), g for group, and o for
other.
To allow yourself to execute a file that you own named myfile, enter:
chmod u+x myfile
To allow anyone who has access to the directory in which myfile is stored to read or execute myfile, enter:
chmod o+rx myfile
You can view the permission settings of a file using the ls command, described below.
Note: Be careful with the chmod command. If you tamper with the directory permissions of your home
directory, for example, you could lock yourself out or allow others unrestricted access to your account and its
contents.
For more detailed information, see the Knowledge Base document In Unix, how do I change the permissions
for a file?

cp
This command copies a file, preserving the original and creating an identical copy. If you already have a file
with the new name, cp will overwrite and destroy the duplicate. For this reason, it's safest to always add -i
after the cp command, to force the system to ask for your approval before it destroys any files. The general
syntax for cp is:
cp -i oldfile newfile
To copy a file named meeting1 in the directory /home/dvader/notes to your current directory, enter:
cp -i /home/dvader/notes/meeting1 .
The . (period) indicates the current directory as destination, and the -i ensures that if there is another file
named meeting1 in the current directory, you will not overwrite it by accident.
To copy a file named oldfile in the current directory to the new name newfile in the mystuff subdirectory of
your home directory, enter:
cp -i oldfile ~/mystuff/newfile
The ~ character (tilde) is interpreted as the path of your home directory.
Note: You must have permission to read a file in order to copy it.

date
The date command displays the current day, date, time, and year.
To see this information, enter:
date
df
This command reports file system disk usage, (i.e., the amount of space taken up on mounted file systems).
For each mounted file system, df reports the file system device, the number of blocks used, the number of
blocks available, and the directory where the file system is mounted.
To find out how much disk space is used on each file system, enter the following command:
df
If the df command is not configured to show blocks in kilobytes by default, you can issue the following
command:
df -k
du
This command reports disk usage (i.e., the amount of space taken up by a group of files). The du command
descends all subdirectories from the directory in which you enter the command, reporting the size of their
contents, and finally reporting a total size for all the files it finds.
To find out how much disk space your files take up, switch to your home directory with the cd command, and
enter:
du
The numbers reported are the sizes of the files; on different systems, these sizes will be in units of either 512
byte blocks or kilobytes. To learn which is the case, use the man command, described below. On most
systems, du -k will give sizes in kilobytes.

find
The find command lists all of the files within a directory and its subdirectories that match a set of conditions.
This command is most commonly used to find all of the files that have a certain name.
To find all of the files named myfile.txt in your current directory and all of its subdirectories, enter:
find . -name myfile.txt -print
To look in your current directory and its subdirectories for all of the files that end in the extension .txt , enter:
find . -name "*.txt" -print
In these examples, the . (period) represents your current directory. It can be replaced by the full pathname
of another directory to search. For instance, to search for files named myfile.txt in the directory
/home/user/myusername and its subdirectories, enter:
find /home/user/myusername/ -name myfile.txt -print
On some systems, omitting the final / (slash) after the directory name can cause find to fail to return any
results.
As a shortcut for searching in your home directory, enter:
find "$HOME/" -name myfile.txt -print
For more detailed information, see the Knowledge Base document In Unix, what is the find command, and how
do I use it to search through directories for files?

jobs
This command reports any programs that you suspended and still have running or waiting in the background
(if you had pressed Ctrl-z to suspend an editing session, for example). For a list of suspended jobs, enter:
jobs
Each job will be listed with a number; to resume a job, enter % (percent sign) followed by the number of the
job. To restart job number two, for example, enter:
%2
This command is only available in the csh, bash, tcsh, and ksh shells.

kill
Use this command as a last resort to destroy any jobs or programs that you suspended and are unable to
restart. Use the jobs command to see a list of suspended jobs. To kill suspended job number three, for
example, enter:
kill %3
Now check the jobs command again. If the job has not been cancelled, harsher measures may be necessary.
Enter:
kill -9 %3
less and more
Both less and more display the contents of a file one screen at a time, waiting for you to press the Spacebar
between screens. This lets you read text without it scrolling quickly off your screen. The less utility is generally
more flexible and powerful than more, but more is available on all Unix systems while less may not be.
To read the contents of a file named textfile in the current directory, enter:
less textfile
The less utility is often used for reading the output of other commands. For example, to read the output of the
ls command one screen at a time, enter:
ls -la | less
In both examples, you could substitute more for less with similar results. To exit either less or more, press q .
To exit less after viewing the file, press q .
Note: Do not use less or more with executables (binary files), such as output files produced by compilers.
Doing so will display garbage and may lock up your terminal.

lpr and lp
These commands print a file on a printer connected to the computer network. The lpr command is used on
BSD systems, and the lp command is used in System V. Both commands may be used on the UITS systems.
To print a file named myfile on a printer named lp1 with lpr, enter:
lpr -Plp1 myfile
To print the same file to the same printer with lp, enter:
lp -dlp1 myfile
Note: Do not print to a printer whose name or location is unfamiliar to you.
For more detailed information, see the Knowledge Base document In Unix, how do I print files and list or
remove print jobs?

ls
This command will list the files stored in a directory. To see a brief, multi-column list of the files in the current
directory, enter:
ls
To also see "dot" files (configuration files that begin with a period, such as .login ), enter:
ls -a
To see the file permissions, owners, and sizes of all files, enter:
ls -la
If the listing is long and scrolls off your screen before you can read it, combine ls with the less utility, for
example:
ls -la | less
For more detailed information, see the Knowledge Base document In Unix, how do I list the files in a
directory?
man
This command displays the manual page for a particular command. If you are unsure how to use a command
or want to find out all its options, you might want to try using man to view the manual page.
For example, to learn more about the ls command, enter:
man ls
To learn more about man, enter:
man man
If you are not sure of the exact command name, you can use man with the -k option to help you find the
command you need. To see one line summaries of each reference page that contains the keyword you specify,
enter:
man -k keyword
Replace keyword in the above example with the keyword which you want to reference. Also see the Knowledge
Base document In Unix, what is the man command, and how do I use it to read manual pages?

mkdir
This command will make a new subdirectory.
To create a subdirectory named mystuff in the current directory, enter:
mkdir mystuff
To create a subdirectory named morestuff in the existing directory named /tmp, enter:
mkdir /tmp/morestuff
Note: To make a subdirectory in a particular directory, you must have permission to write to that directory.
mv
This command will move a file. You can use mv not only to change the directory location of a file, but also to
rename files. Unlike the cp command, mv will not preserve the original file.
Note: As with the cp command, you should always use -i to make sure you do not overwrite an existing file.
To rename a file named oldname in the current directory to the new name newname, enter:
mv -i oldname newname
To move a file named hw1 from a subdirectory named newhw to another subdirectory named oldhw (both
subdirectories of the current directory), enter:
mv -i newhw/hw1 oldhw
If, in this last operation, you also wanted to give the file a new name, such as firsthw, you would enter:
mv -i newhw/hw1 oldhw/firsthw
ps
The ps command displays information about programs (i.e., processes) that are currently running. Entered
without arguments, it lists basic information about interactive processes you own. However, it also has many
options for determining what processes to display, as well as the amount of information about each. Like lp
and lpr, the options available differ between BSD and System V implementations. For example, to view
detailed information about all running processes, in a BSD system, you would use ps with the following
arguments:
ps -alxww
To display similar information in System V, use the arguments:
ps -elf
For more information about ps refer to the ps man page on your system. Also see the Knowledge Base
document In Unix, what do the output fields of the ps command mean?

pwd
This command reports the current directory path. Enter the command by itself:
pwd
For more detailed information, see the Knowledge Base document In Unix, how do I determine my current
working directory?

rm
This command will remove (destroy) a file. You should enter this command with the -i option, so that you'll
be asked to confirm each file deletion. To remove a file named junk, enter:
rm -i junk
Note: Using rm will remove a file permanently, so be sure you really want to delete a file before you use rm.
To remove a non-empty subdirectory, rm accepts the -r option. On most systems this will prompt you to
confirm the removal of each file. This behavior can be prevented by adding the -f option. To remove an entire
subdirectory named oldstuff and all of its contents, enter:
rm -rf oldstuff
Note: Using this command will cause rm to descend into each subdirectory within the specified subdirectory
and remove all files without prompting you. Use this command with caution, as it is very easy to accidently
delete important files. As a precaution, use the ls command to list the files within the subdirectory you wish to
remove. To browse through a subdirectory named oldstuff, enter:
ls -R oldstuff | less
rmdir
This command will remove a subdirectory. To remove a subdirectory named oldstuff, enter:
rmdir oldstuff
Note: The directory you specify for removal must be empty. To clean it out, switch to the directory and use
the ls and rm commands to inspect and delete files.

set
This command displays or changes various settings and options associated with your Unix session.
To see the status of all settings, enter the command without options:
set
If the output scrolls off your screen, combine set with less:
set | less
The syntax used for changing settings is different for the various kinds of Unix shells; see the man entries for
set and the references listed at the end of this document for more information.
vi
This command starts the vi text editor. To edit a file named myfile in the current directory, enter:
vi myfile
The vi editor works fairly differently from other text editors. If you have not used it before, you should
probably look at a tutorial, such as the Knowledge Base document How do I use the vi text editor? Another
helpful document for getting started with vi is A quick reference list of vi editor commands.
The very least you need to know to start using vi is that in order to enter text, you need to switch the program
from command mode to insert mode by pressing i . To navigate around the document with the cursor keys,
you must switch back to command mode by pressing Esc. To execute any of the following commands, you
must switch from command mode to ex mode by pressing : (the colon key): Enter w to save; wq to save
and quit; q! to quit without saving.
w and who
The w and who commands are similar programs that list all users logged into the computer. If you use w, you
also get a list of what they are doing. If you use who, you also get the IP numbers or computer names of the
terminals they are using.
In Unix, how do I kill another login session remotely?
You can kill a Unix login session remotely by sending a hangup signal (SIGHUP) to the process running the
login session. To do this, follow the steps below:
1. Identify the shell you want to kill. To determine your current tty, from your Unix shell prompt, enter:
tty
2. To show all of your running processes, enter: ps -fu username Replace username with your
username.

3. You should see something like this: PID TT STAT TIME COMMAND

4. 13964 v5 I 0:00 elm

5. 13126 ue S 0:00 -bash (bash)

6. 13133 ue R 0:00 ps x

7. 13335 v5 S 0:00 -bash (bash)

In the first column, "PID" stands for "process ID". The second column shows the tty to which your
processes are connected. The dash ( - ) before a process name shows that the process is a login shell.
8. To remove the remote shell, look for the processes with a dash and choose the process number that is
not for your current tty. Then issue the following command: kill -HUP processid Replace processid
with the process ID number you identified.
When you send a SIGHUP (by entering kill -HUP or kill -1) to a login shell, all the processes that were started
in the shell will be killed as well (unless they were in the background). SIGHUP is good because it allows
applications like Elm and Emacs to exit gracefully, leaving your files intact.
Note: You cannot kill processes that are running on a computer different from the one you are logged into.
For example, at Indiana University you can't kill processes running on Libra from your Steel account. This rule
extends to clusters of Unix systems as well (e.g., Steel, Libra, da Vinci, and Nations). This means that you
cannot kill a process running on the Steel node Steel1 if you are logged into Steel2.

In Unix, how do I cancel a batch job?

In Unix, if you scheduled a job with at or batch, you can cancel it at the Unix prompt, by entering: at -r
<jobnum>
Replace <jobnum> with the number of the job that at or batch reported when you submitted the job. On some
systems, you may use atrm instead of at -r .
If you don't remember the job number, you can get a listing of your jobs by entering: at -l
Each job will be listed with its job number queue and the time it was originally scheduled to execute.
On some systems, the atq command is available to list all the jobs on the system. To use this command, at
the Unix prompt, enter: atq
If your job is already running, you will need to find the process ID and kill it. On System V implementations
(including all UITS central systems at Indiana University), list all running processes by entering: ps -fu
username Replace username with your username. The equivalent BSD command is: ps x
Once you have the process ID, enter: kill <process ID> Replace <process ID> with the process ID. If it still
will not terminate, try entering: kill -9 <process ID>
Note: Other scheduling programs, such as NQS and LoadLeveler, work differently and are not controlled by
the same methods. For more information, see the appropriate man pages.
In Unix, how do I check the CPU usage of a job?
Note: If you are concerned about slowing the system down, you can use the nice command to lower your
program's priority. For more information about the nice command, at the Unix prompt, enter: man nice
In Unix, you can see CPU usage on a job that is running in a number of ways, as described below:
The time command
If you want to see a grand total of CPU time for a program when it finishes running, you can use the time
command. At the Unix prompt, enter: time myprog Replace myprog with the name of the program you are
running. The following is an output example for users in the csh or tcsh shells: 1.406u 0.042s 0:04.96 29.0%
2+5k 0+1io 0pf+0w The program myprog used 1.406 seconds of user time, 0.042 seconds of system time,
and 4.96 seconds of real time. The sum of the user and system times is the total CPU time of the process. The
percentage (29.0%) indicates the percentage of the CPU's time that the process used while it ran. The output
will appear in a slightly different format when using sh, ksh, or bash, since the time command is not built into
those shells.
The ps command
You can also use the Unix command ps. At the Unix prompt, enter: ps -u username Replace username
with your username. You will see something like the following: PID TTY TIME COMMAND

10005 ? 3:03 a.out

15852 rb 0:01 -tcsh (tcsh)

24980 rb 0:00 sh -c /usr/local/bin/emacs /usr1/mmouse/snd.18106

24981 rb 0:02 /usr/local/bin/emacs /usr1/mmouse/snd.18106

22311 rf 0:22 elm In this example, the "TIME" column shows that the process running Elm has used 22
CPU seconds.
The top command
You may also use the top command. At the Unix prompt, enter: top You will see something similar to
the following: PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND

28000 mmouse 96 4 276K 144K run 292:20 16.80% 16.80% desert.exe

27999 mmouse 96 4 276K 144K run 292:42 16.02% 16.02% denver.exe

19004 goofy 96 4 428K 160K run 357:11 15.63% 15.63% diskr

User mmouse is at the top of the list, and the "TIME" column shows that the program desert.exe has used 292
minutes and 20 seconds of CPU time. This is the most interactive way to see CPU usage.
How do I run a Unix process in the background?
In Unix, a background process executes independently of the shell, leaving the terminal free for other work. To
run a process in the background, include an & (an ampersand) at the end of the command you use to run the
job. Following are some examples:
 • To run the count program, which will display the process identification number of the job, enter: count
&
• To check the status of your job, enter: ps
• You can bring a background process to the foreground by entering: fg
• If you have more than one job suspended in the background, enter: fg %# Replace # with the job
number, as shown in the first column of the output of the jobs command.

•
You can kill a background process by entering: kill PID Replace PID with the process ID of the job. If
that fails, enter the following: kill -KILL PID
• To determine a job's PID, enter: jobs -l
• If you are using sh, ksh, bash, or zsh, you may prevent background processes from sending error
messages to the terminal. Redirect the output to /dev/null using the following syntax: count 2>
/dev/null &
Unix job control command list
The following table lists the basic Unix job control commands:

Command Explanation Example

& Run the command in the background % long_cmd &
Ctrl-z Stop the foreground process [Ctrl-z]
Stopped
jobs List background processes % jobs
[1] - Stopped vi
[2] - big_job &
%n Refers to the background number n % fg %1
%?str Refers to the background job containing str % fg %?ls
bg Restart a stopped background process % bg
[2] big_job &
fg Bring a background process to the foreground % fg %1
kill Kill a process % kill %2
~ Ctrl-z Suspend an rlogin or ssh session host2>~[Ctrl-z]
Stopped
host1>
~~ Ctrl-z Suspend a second level rlogin or ssh session host3>~~[Ctrl-z]
Stopped
host2>
This table is adapted from Essential System Administration, by Aeleen Frisch, copyright 1995, O'Reilly &
Associates, Inc.
Note: For security reasons, rlogin is not available on UITS computers at Indiana University.

In UNIX, what do the output fields of the ps command mean?

The ps command varies significantly among Unix implementations. Each vendor incorporates its own flags and
outputs the results differently. However, most ps variants are rooted enough in either the System V or BSD
syntax that entering ps -elf (System V) or ps alx (BSD) will produce something like the following:
F S UID ID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME COMD
1 R obiwan 792 779 22 183 20 10ec5f80 29 - 12:52:24 pts/2 0:00 ps -elf
1 S root 24621 560 0 154 20 13603f80 11 4697c0 Jun 16 ttyp2 0:00 telnetd
1 S dvader 1162 1153 0 154 20 110a1f80 77 452be4 11:25:41 pts/3 0:00 ssh deathstar
This particular example is from HP-UX, whose output is basically vanilla System V. The following table
describes the meanings of the columns that commonly appear in ps outputs. No version of ps will display all of
these fields, however.

Column Header Contents

%CPU How much of the CPU the process is using
%MEM How much memory the process is using
ADDR Memory address of the process
C or CP CPU usage and scheduling information
COMMAND* Name of the process, including arguments, if any
NI nice value
F Flags
PID Process ID number
PPID ID number of the process's parent process
PRI Priority of the process
RSS Real memory usage
S or STAT Process status code
START or STIME Time when the process started
SZ Virtual memory usage
TIME Total CPU usage
TT or TTY Terminal associated with the process
UID or USER Username of the process's owner
WCHAN Memory address of the event the process is waiting for
*
= Often abbreviated
For information specific to your Unix implementation, consult the ps man page.