The commandline

Navigating around

Print current working directory:

pwd

List files and folders current directory:

ls

List a directory with more details:

ls -l

Above -l stands for long. In some shells, the above command is aliased to ll.

List hidden (all) files:

ls -la

You can list not only current directory but another directory:

ls sub_directory

Common ls options:

Option	Description
-a	List all files
-h	Display file sizes in more human readable format instead of bytes
-l	Long format
-r	Reverse order
-S	Sort by size
-t	Sort by time

Your current directory is denoted by ., and parent directory by ..

There is a nice utility tree which displays the tree structure of a directory:

tree
tree -aCd -L 2 /

Change directory:

cd home
cd ..

Go to previous working directory:

cd -

Go to home directory (any of the below):

cd
cd ~

Files and folders

Determine a file content with file command:

file profile.jpg

Copy file_original to file_final:

cp file_original file_final

Move file or directory:

mv file_source file_destination

Copy a directory along with the contents inside recursively:

cp -r source_path destination_path

Sometimes we need to copy files along with permissions, timestamps, and other attributes, use the -p flag:

cp -rp source_path destination_path

Create new directory:

mkdir new_dir

Delete a directory:

rm -rf my_dir

Delete an empty directory:

rmdir empty_dir
# or use its alias
rd empty_dir

tip

One of the fastest way to delete a large directory:

mkdir empty && \
  rsync -r --delete empty/ large_dir && \
  rm -r large_dir empty

Disk usage, find folder size:

du -hs
du -hs path/

# list all folders with wild card
du -hs *

# list folder sizes and sort
du -hs * | sort -n

# total usage line in the end
du -hsc *

# list sizes by controlling depth
du -d 2 -h

Display string/variable:

echo string
echo $variable

# with variable expansion
echo ${variable}

# write/rewrite to a file
echo string > file

# append to an existing file
echo string >> existing-file

Print (concatenate the contents to the standard output) the contents of a file:

cat file.txt

Print only the head (default first 10 lines) of a file:

head file.txt

You can print particular number of lines:

head -20 file.txt

Print as much as fits in the shell window:

more file.txt

See the end of a file:

tail file.txt
tail -20 file.txt

Print the contents to a file:

cat file.txt > file-copy.txt

This seems same as copy. But you can use this to combine multiple files:

cat file.txt, file2.txt > file-copy.txt

Instead of overwriting file, you can append:

cat file3.txt >> file-copy.txt

You can use paste to combine files as well:

paste file* > file-all.txt

Text redirect with sudo:

echo "something" | sudo tee /ect/something.config

If you need to append instead of overwriting:

echo "something" | sudo tee -a /ect/something.config

Redirect multiple lines with cat:

# cat << EOF > test.txt
cat > test.txt << EOF
Hello World!
Have a good day.
EOF

Redirect multiple lines with sudo privilege:

cat << EOF | sudo tee /etc/jupyter/jupyter_server_config.py > /dev/null
c.ServerApp.ip = '0.0.0.0'
c.ServerApp.allow_root = True
c.ServerApp.open_browser = False
EOF

Look for lines containing certain word. This will list the lines that contain the word file in the cat manual:

man cat | grep file

Working with large files:

less file.txt
man top | less
cat file.txt | less

Common less keyboard shortcuts:

Shortcut	Description
space	Scroll forward one page
b	Scroll back one page
G	Move to the end of file
g	Move to the beginning of file
/char	Search forward
n	Search for next occurrence
q	Quit

Redirecting output:

# throw away output
npm install &> /dev/null

Writing stdout and stderr to a file

npm install &> log.txt

# or
npm install >log.txt 2>&1

# write to separate files
npm install > result_log.txt 2> error_log.txt

# throw away stderr
npm install > result_log.txt 2> /dev/null

npm install 2>&1 > result_log.txt

Create symbolic links:

ln -s /path/to/file.txt /path/to/link.txt

Bash expansion: If you need to create/copy/move/delete a series of files:

touch file_{01..05}.txt

It will create file_01.txt, file_02.txt and so on.

Recall previous command:

apt update
sudo !!

tar

Create a tar archive of files:

tar -czvf files.tar.gz file1.pdf file2.pdf

Create a tar archive of a directory:

tar -czvf files.tar.gz /home/files

Untar an archive:

tar -xzvf files.tar.gz

Extract on a different directory:

tar -xzvf files.tar.gz -C /tmp/

List contents of a tar file:

tar -tf files.tar.gz

zip

If you have got a zip file to unzip. You need to install unzip.

sudo apt install zip unzip
sudo dnf install zip unzip

Then:

unzip filename.zip
unzip filename.zip -d /home/destination

# see contents of zip file without unzipping
unzip -l filename.zip

Zip file(s) or folders:

zip output.zip file1.txt file2.md
zip -r output.zip source_dir
zip -rq output.zip source_dir

Encrypt ZIP files using 7-zip. The ZIP files are compatible with Windows (use 7-zip, or WinZip/WinRAR) and macOS (use Keka).

# install 7-zip in ubuntu/debian
sudo apt update
sudo apt install 7zip

# encrypt and add to zip archive
7zz a archive.zip file1.txt file2.pdf -mem=AES256 -p -tzip

# encrypt directory
7zz a -mem=AES256 -p -tzip archive.zip dir

7zz a -mem=AES256 -p -tzip archive.zip -r dir/*

7zz a -tzip archive.zip -r src/*.c src/*.h

# alternatively you can provide password with the `-p${passwd}` flag

# list contents of archive
7zz l archive.zip

# extract files
7zz x archive.zip

# set output directory
7zz x archive.zip -o./new-dir

caution

The above method does not encrypt the file/directory names while using ZIP format. If you want to encrypt archive headers, use -mhe switch with other compression methods such as 7z. Find more details here and here.

Verify the checksum

md5 path/file.iso
shasum -a 256 path/file.iso
shasum -a 512 path/file.iso

find

Search for files using find command:

find /home/pranab -type f -name profile.jpg

# find in several directories
find /home/pranab /users/pranab -type f -name profile.jpg

# case insensitive search
find /home/pranab -type f -iname profile.jpg

# by filetype
find /home/pranab -type f -name *.pdf

Useful -type flags: f: regular files, d: directory, l: symbolic link.

Find by file permission:

find . -perm 0600  -print
find . -perm 0644  -type f -name tests.py -print

# without permission
find . -type f ! -perm 777

# find all executables
find . -perm /a=x

Find by file size:

find /tmp -type f -size 1024c

# less than 1 megabyte
find . -type f -size -1M

# bigger than 1 kilobyte
find . -type f -size +1k

# between 512 MB and 1 GB
find . -type f -size +512M -size -1G

Find files based on modification time

# find files modified in last 7 days
find . -name "*.md" -mtime 7

# modified in last hour
find . -name "*.md" -mmin 60

# last modified more than 7 days before
find . -name "*.md" -mtime +7 -daystart

Find and delete file:

find . -type f -name .DS_Store -delete

danger

Run above command with caution. First run the command without the -delete flag to be sure what your are deleting. If the expression prior to -delete is what you did not mean, it could delete something that you did not want.

Look for multiple filetype at once:

find . -type f \( -name \*.jpg -o -name \*.png \)

grep

Find specific string/words in a file or set of files.

grep estimation output.txt
grep "energy estimation" output.txt

# find in multiple files
grep estimation output.txt output2.txt

# find in all files in a directory
grep estimation *

# ignore case
grep -i estimation *

# search subdirectories
grep -r estimation *

# list of matching files
grep -l estimation *

# count number of matches
grep -c estimation *

# search for multiple strings
grep -e estimation -e "total energy" *

sed

Use stream editor to find and replace text.

# substitute first string occurrence in every line
sed s/pattern/replace_string/ file.txt

# substitute all string occurrences in every line
sed s/pattern/replace_string/g file.txt

# substitute al string occurrences in range of lines
sed 1,5s/pattern/replace_string/g file.txt

# case insensitive search
sed s/pattern/replace_string/gI file.txt

# write the output to a file
sed s/pattern/replace_string/gI file.txt > newfile.txt

# save changes to the same file
sed -i.bak s/pattern/replace_string/gI file.txt

In the last command it will create a backup of the original file named file.txt.bak.

wget

Download a file using wget:

wget https://example.com/pic.jpg

Download and save to a path/name:

wget https://example.com/pic.jpg -O images/profile.jpg

Continue with interrupted download:

wget -c http://example.com/file.ext

Mirror a website locally using wget:

wget --mirror \
  --convert-links \
  --html-extension \
  --wait=1 \
  --user-agent="Mozilla/5.0 (Linux x86_64; rv:97.0) Gecko/20100101 Firefox/97.0" \
  -o log \
  http://example.com

--convert-links will convert the links suitable for viewing with local paths; --html-extensions will handle if there are no extension at the end of URLs; --wait to specify waiting time between successive retrievals in seconds.

ssh

ssh can be used to connect to remote machines. First make sure you have the ssh installed in your machines:

sudo apt install --no-install-recommends openssh-server

You need to know the IP address or hostname of remote machine:

ip addr

Connecting to a ssh server:

ssh <username>@<xxx.xx.xx.xxx> -p <yyyy>

where xxx.xx.xx.xxx is the IP address and yyyy is the port number.

You can disconnect from the remote ssh by pressing [control]+[d] or typing exit on the terminal.

Running ssh server on Windows Subsystem for Linux. Open the following file:

sudo vi /etc/ssh/sshd_config

Change the port to anything above 1000 (default 22 might conflict with host Windows system if you are using WSL). Change PasswordAuthentication yes Add AllowUsers <your username>

Save and exit. Following commands should be self explanatory:

service ssh status
sudo service ssh start
sudo service ssh stop
sudo service ssh --full-restart

Port forwarding. Let's say you are working on a python jupyter notebook on the remote server via ssh and you want to see the notebook in your local browser. You can forward the specific port to your local machine:

ssh -L 8888:127.0.0.1:8888 pranab@xxx.xx.x.xxx -p yyyy -i ~/.ssh/id_rsa

Now you can see the notebook in your client machine.

Adding SSH key-pair to connect to a server. Generate the key-pair in your client computer:

ssh-keygen

# generate key with specific algorithm
ssh-keygen -t ed25519 -C "email@example.com"

# for legacy systems use 4096 bit RSA
ssh-keygen -t rsa -b 4096 -C “email@example.com”

# find bit length of an existing key
ssh-keygen -l -f ~/.ssh/id_ed25519.pub

# add or change password of an existing key
ssh-keygen -p -f ~/.ssh/id_ed25519

You may leave everything default and press enter. It will generate public and private key pair inside ~/.ssh directory. Now we need to copy the public key to the host machine. Connect to remote host using password. Create a .ssh directory in the remote home directory. You can print the public key in your terminal cat id_rsa.pub and copy it to clipboard.

ssh <username>@<xxx.xx.xx.xxx> -p <yyyy>
mkdir .ssh
vi .ssh/authorized_keys

Paste the public key, save and exit. Then open /etc/ssh/sshd_config with sudo vi and uncomment the line PubkeyAuthentication yes. You may choose to set PasswordAuthentication no if you want. Set the following permissions:

sudo chmod 700 .ssh
sudo chmod 600 .ssh/authorized_keys

Restart the ssh service on remote host

sudo service ssh restart

Now you can connect from your local computer using:

ssh pranab@xxx.xx.x.xxx -p yyyy -i ~/.ssh/id_rsa

Copying files via scp from remote to local or vice-versa:

scp remote_username@10.10.0.2:/remote/file.txt /local/directory
scp file.txt remote_username@10.10.0.2:/remote/directory

You can use scp to transfer files as well using ssh-keys:

scp -i ~/.ssh/id_rsa remote_username@10.10.0.2:/remote/file.txt /local/directory
scp -P yyyy -i ~/.ssh/id_rsa file.txt remote_username@10.10.0.2:remote/directory

We can secure copy a directory using -r flag:

scp -i "ssh_key.pem" -r ~/Documents/mydir user@domain.com:Downloads/mydir

diff

Compare two files:

diff file.txt file_new.txt

Compare two folders:

diff -rq folder folder_new

Here the q flag is for the brief mode, otherwise it would not only show which files are different but also show line-by-line differences.

openssl

There are various encription options available in openssl. In the above, we have used AES-256 (Advanced Encryption Standard with 256 bit key length) symmetric encryption in CBC (Cipher Blocker Chaining) mode, used salt hashing, pbkdf2 (Password-Based Key Derivation Function 2) key stretching, and armor output (-a). Check openssl documentation for more details.

Some good set of options:

# to encrypt
openssl enc -e -aes-256-cbc \
  -salt \
  -iter 1000000 \
  -md sha512 \
  -base64 \
  -in somefile.txt \
  -out somefile.enc

# to decrypt
openssl enc -d -aes-256-cbc \
  -salt \
  -iter 1000000 \
  -md sha512 \
  -base64 \
  -in somefile.enc \
  -out somefile.txt

note

It is very important to write down which encryption options you use in order to decrypt them. Default options may not be the best choice.

-salt flag is very crucial, it is used by default while encrypting using openssl. Use of -iter also enables -pbkdf2 for key stretching.

The ASCII armor (base64) output will increase the file size. If you don't need to copy and pase the output (say, you need to send via an email message), it is better to leave them as binary.

If you are going to compress your data, you should do it before encrypting. After encryption, it is basically random gibberish which won't benefit from compression algorithms (e.g., dictionary based compression).

You can also use GPG program to perform symmetric encryption.

For on the fly encryption and decryption with terminal input and output:

openssl enc -e -aes-256-cbc \
  -salt \
  -iter 1000000 \
  -md sha512 \
  -base64 <<< "Secret message."

openssl enc -d -aes-256-cbc \
  -salt \
  -iter 1000000 \
  -md sha512 \
  -base64 <<< "U2FsdGVkX18Y9d6MT+d8kbrmfsgd1j/vnNlZ5T7LQ6I="

If you are on Windows, you can get OpenSSL by installing git. You can consider having the portable git version if you don't need to install git in your computer. Run using full path to executable: 'C:\Program Files\Git\usr\bin\openssl.exe' if the bin directory is not added to PATH environment variable.

File permissions

When we list a file using ls -la or similar, we see permissions strings. What does they mean?

r → read
w → write
x → execute

and they are listed for user, group, and others. We can set the file permissions by using numeric conventions. r=4 (binary 100), w=2 (010), and x=1 (001). Say we want to set rwx for user, rx for group and only x for others, we can set that by:

chmod 751 file.sh

Alternatively, in symbolic notation, we can assign or add or remove permissions:

chmod u=rwx, g=rx, o= file.sh
chmod u+x file.sh
chmod g-x file.sh

We can change file permissions recursively as well:

chmod -R ugo-x /home/pranab
chmod -R a-x /home/pranab

Make a script file executable:

1. Create a file filename.sh
2. Write down your script
3. Make the file executable: chmod +x filename.sh
4. Run the script: ./filename.sh

rsync

It can be used to mirror/copy files and folders in a local machine as well as to / from remote machines. It is a very powerful tool with lots of options and customizations. Here are few common ones:

Copy / sync files locally:

rsync -avh backup.tar /tmp/backups/

Copy folders locally:

rsync -avh /root/rpmpkgs /tmp/backups/
rsync -avh /root/rpmpkgs/ /tmp/backups/rpmpkgs

Copy files over ssh:

rsync -avzhe "ssh -i ~/.ssh/id_rsa -p yyyy" pranab@xxx.xx.x.xxx:file.txt ~/Desktop/
rsync -avzhe "ssh -i ~/.ssh/id_rsa -p yyyy" ~/Desktop/file.txt pranab@xxx.xx.x.xxx:docs

By default rsync won't remove deleted files from the destination. If you want to delete:

rsync -avh --delete /home/file.txt /usr/desktop/

While using the --delete flag, it is a good idea to first run the command with --dry-run flag to see the changes before actually deleting the files. The --dry-run flag let you see the output without actually applying the command.

Exclude a file or directory from syncing:

rsync -avh --delete --exclude .git /home/my-project/ /mnt/d/my-project
rsync -avh --delete --exclude *.iso /home/my-project/ /mnt/d/my-project
rsync -avh --max-size=500m /home/my-project/ /mnt/d/my-project
rsync -avh --exclude={'*.log','dir-x','dir-y'} /home/my-project/ /mnt/d/my-project
rsync -avh --exclude-from='exclude.txt' /home/my-project/ /mnt/d/my-project

Where exclude.txt contains list of files and directories:

exclude.txt

.gitignore
.git
*.log
test*

Commonly used flags:

Flags	Description
-v	verbose
-r	copies data recursively, doesn’t preserve timestamps and permission
-a	archive mode, copies recursively and preserves ownerships and timestamps
-z	compress data while transferring
-h	human-readable
-e	with protocol option.

If you are looking for more advanced and secure backup options, check out https://borgbackup.readthedocs.io/en/stable/

Processes

See processes in the current shell

ps

Show all processes:

ps -ef

Show all processes of all users

ps -aux
ps -aux | head -10

You can exit by typing q. You can press k and provide PID to kill a process.

You can kill a process by:

kill -SIGKILL <pid>

On the command line, you can terminate a command by CTRL+C or suspend by CTRL+Z.

View process tree:

ps tree

Check load balance:

top

tmux

tmux has many use case scenarios, one them is running something on a remote machine without interruption. Say you are working on a remote machine over ssh, and you need to update your remote machine. The updating process can be catastrophic if it is interrupted, it can leave your system on a broken state. Therefore it is a good idea to run the process using tmux so that even if you are disconnected from the ssh session, the process can continue.

Installing tmux:

sudo apt install tmux

Launch a tmux window:

tmux

Run a process/command there. Now we can detach from tmux using keyboard shortcut: first send the tmux bind-key [control] + [b] followed by [d]. Now we get back our main terminal. We can create multiple tmux sessions. We can list tmux sessions by:

tmux ls

We can attach particular session:

tmux attach -c 2

Or attach the last working session or if there is only one session:

tmux attach

When we are in one of tmux window, we can use [control] + [b] followed by [w] for tmux window list.

We can exit the tmux session by typing exit on the tmux window. Or when we are in main terminal:

tmux kill-session -t 0

Terminal shortcuts

Shortcuts	Description
CTRL + A	Move to the beginning of a line
CTRL + E	Move to the end of a line
CTRL + K	Kill/delete/cut from the cursor to the end
CTRL + U	Delete from the cursor to the start of the line
Ctrl + X + Backspace	removes all the text from the cursor to the beginning
CTRL + W	Delete from cursor to the start of word
CTRL + Y	Undo last deletion
ALT + Left arrow	Move left one word
ALT + Right arrow	Move right one word
CTRL + L	Clear screen
CTRL + C	Terminate the foreground process
CTRL + Z	Suspend the foreground process
CTRL + D	Exit shell.

HereDoc

[COMMAND] <<[-] 'DELIMITER'
  Line 1
  Line 2
  ...
DELIMITER

• COMMAND is optional. Works for any command that accepts redirection.
• << is the redirection operator for forwarding a HereDoc to the COMMAND.
• - is an optional parameter for tab suppression. Tab (not spaces) can be used for indentation.
• DELIMITER in the first line defines a HereDoc delimiter token. END, EOT, and EOF are most common, but any multicharacter word that won't appear in the body works. Omit single quotes on the first line to allow command and variable expansion. The DELIMITER in the last line indicates the end of a HereDoc. Use the same word from the first line without the leading whitespaces.

Learn more

• Art of commandline
• IBM Linux tutorials
• The Linux command line: a complete introduction by William E. Shotts.
• Managing SSH keys