Basic Linux commands
Using the computer clusters requires familiarity with basic Linux commmands. After logging into your cluster account, you are facing a Linux shell, which takes commands and performs certain tasks. If you have been used to a graphical point-click interface this might seem daunting. However once you get familiar with the commands you’ll find it extremely powerful and convenient. I think there are only 20 or so important commands one needs to know in order to perform most of the tasks. Here is a list of them with VERY brief explanations. Each of these commands takes tons of options, but only a few of them are frequently used. For more detailed explanations please refer to the manual pages.
man: to get help for a command. This is short for
manual. For example to get the complete help page for
man lsat the command line.
ls: get a list of the directory contents. Most common option is
ls -l, which list the contents in long format (one line per item showing file modes, owner, time accessed, size, etc). Other options I often use are
ls -rtl, which list files in long format and sorted by time with the latest file at the bottom, and
ls -rSl, which list files in long format and sorted by size with the largest file at the bottom.
cp: copy files. Typical usage is
cp source_dir/source_file target_dir/target_file. An frequently used option is
-r, which copies directories. For example,
cp -r * target_dir/</tt>copies all contents including sub directories into folder named
scp: secure copy files from/to a remote file system (such as the cluster). For example,
scp source_file email@example.com:~copies
source_fileto the home directory of user
scp -r firstname.lastname@example.org:~/* target_dircopies everything in
hwu30’s home directory at
cluster.sph.emory.eduto local directory
target_dir. Note that
cpshare many parameters.
mv: move/rename files. Typical usage is
mv source_dir/source_file target_dir/target_fileDifferences between
mvare (1) after
mvthe source file will be deleted; and (2) time stamp for the source file will be perserved in the target file.
rm: remove (delete) files:
rm target_dir/target_file. This should be used in caution because usually one cannot undelete a file or folder once they are rm’ed. Frequently used option is “
-r” for removing a directory.
mkdir/rmdir: make/remove directories. For example mkdir dir1 makes a subdirectory called “dir1” in current directory. rmdir can only delete an empty directory. To delete a directory with contents use “rm -r”.
cd: change directory. For example: cd dir1/subdir2/subsubdir3/.
pwd: shows the path of current directory.
- more/less: to view a text file. This is extremely useful in viewing
large text files for that they don’t open and load the whole file into memory,
which makes it very fast. more and less are pretty similar,
you can pick one your like. Within the program some useful commands are:
- space: go one page down.
b: go one page up.
/pattern: search for pattern in the text file.
q: quit to the shell.
head/tail: to show the start or end a few lines of a text file. For example “head -5 file.txt” shows the first 5 lines of file.txt.
wc: counting number of words and lines of a file. For example if “wc file.txt” returns
25700 777244 8974608 file.txtIt means file.txt has 25700 lines, 777244 words and 8974608 bytes.
zip/unzip: To compress/expand files. Typical usage is
zip target.zip *.txt. This compresses all txt files into one zip file called target.zip. To extract them type unzip target.zip.
gzip/gunzip: To compress/expand files. gzip can only compress one file at a time, that’s why it’s always used with tar, which concatenate multiple files into one (without compressing). For example, gzip file.txt generates file.txt.gz and deletes the original file.txt. To decompress, simply type gunzip file.txt.gz.
tar: manipulate tar balls (one big file contains many little files). There’s no compulsive reason of creating a tar if one can use zip in these days. However very often the Linux based software are distributed as .tar.gz files, which is a gzipped tar ball. It’s useful to know how to extract the whole thing, which is tar xvfz software.tar.gz.
du: display disk usage. For example “du -h” shows total disk usage in current directory. If there are many subdirectories it will print out a long list of usages. In this case “du –max-depth=1 -h” is useful, which only shows one level down of the subdirectories.
grep: look for pattern matching from files.
grep apattern *.txtsearches for all txt files in current directory and show the lines containing “apattern”. “grep” is a complicated command taking tons of options. “grep” is a very powerful command (100 times better than the little dog in Windows!) if you know how to use it. For example: grep -r –include=”*.R” apattern * recursively searches all R files under current directory (including all subdirectories) looking for lines with “apattern”. The search pattern takes regular expression.
- find: find file(s) within a directory and all sub-directories. For example, if I want to find all R files in a directory, I can do
find . -name "*.R"
Other useful tips on a linux system
Press “TAB” key to complete the name of a command, a file or a directory when typing. For example is there is a file called “reallylongfilename.txt”. You can type in “really” (or the first a few letters) then press TAB, and the file name will be completed automatically.
Press up/down arrow one can find previously typed commands in terminal.
By default the terminal takes Emacs hotkeys: Ctrl-A moves the cursor to the beginning of line, Ctrl-E moves to the end of the line, Ctrl-R searches commands backward, etc. Once you get used to this you will find it extremely convenient.
Using greater than signs (“>”) can write the standard output (things printed on the screen) to a file. For example, “ls -l > dirls.txt” writes the directory listing to a file named “dirls.txt”. This with less than sign (“«/tt>”) and vertical bars (“|”) are used in the powerful Linux piping, which means the output from one process becomes the input of another one.
Descriptions of some other useful (but a little advanced) commands (xargs, awk, cut, sort, sed) can be found at here.
That’s it! I suggest you buy a Linux cheat shirt, which will greatly help memorizing these commands.