This is a collection of Linux tips accumulated over the years.
Even if you are are a seasoned Linux user, you may still learn one thing or two. For example, do you know how to make bash unsensitive to filename case during TAB completion? Create bookmarks for directories so that you can jump (cd) directly into them? List all bash functions defined in the current environment? Parallelize xargs
? Rename files using bash variable substitution? Search files that have been modified in the last 2 weeks? Encrypt a file on the command line with gpg
? Wake up a remote computer which is powered off? Read on.
When you type command lines, you are sending instructions to a program called the “shell”. Unless otherwise indicated, the tips assume that you are using the Bash shell. You can check which shell is running with the command:
echo $SHELL
For a basic introduction to interacting with the shell, I recommend Learning the shell and The Linux command line by William Shott, or any of the following tutorials:
- UNIX Tutorial for Beginners
- Openclassrooms Initiation à Linux (in French)
- YouTube series on The Linux Command Line
- Classic Shell Scripting book
Bash relies on the readline library to interact with the user. Its behavior can be customized by options in $HOME/.inputrc
. For example, to enable case-insensitive tab-completion:
echo 'set completion-ignore-case On' >> ~/.inputrc
- Ctrl-d : delete character under the cursor
- Ctrl-k : delete everything from the cursor till the end of the line
- Ctrl-u : delete everythin from the start if the line untl the cursor
- Alt-d : delete till the end of the current word
- Ctrl-a : move the cursor to the beginning of the line
- Ctrl-e : move the cursor to the end of the line
- Ctrl-b, Alt-b: move backward, by one char or by one word
- Ctrl-f, Alt-f: move forward, by one char or by one word
- Ctrl-r : interactive backward search in the history of command lines
- Ctrl-p : move to previous command line
- Ctrl-n : move to next command line
If you start a command line with a space, this line will not be saved in the history of commands.
At any time, you can clear the list of commands typed in the current session with history -c
The history of commands lines typed in the current shell is saved on the disk (that is, appended to the file pointed to by $HISTFILE
) only when you exit this shell. But you can write them explicitly at any time using history -w
I recommend to set $HISTSIZE
and $HISTFILESIZE
to large numbers, e.g.:
export HISTSIZE=10000
export HISTFILESIZE=100000
in .bashrc
Check out man readline
for many other keybindings and to customize the behavior of the command lines by setting variables in ~/.inputrc
- Ctrl-c: Terminate (kill) the running job
- Ctrl-z: Stop the running job (you can then put in background with
bg
) - Ctrl-d: Terminate the Shell and close the Terminal
- Ctrl-l: Clear the terminal
- Ctrl+s: Stop all output to the screen. This is particularly useful when running commands with a lot of long, verbose output, but you don’t want to stop the command itself with Ctrl+C.
- Ctrl+q: Resume output to the screen after stopping it with Ctrl+S.
When you type something on a command line and press Return
, you give an order to the shell. The first token is a command and the following tokens on the line are parameters, a.k.a. agurments.
There are different types of commands:
- shell builtins
- user-defined functions
- aliases
- programs (scripts or binary executable files), that is, executable files located somewhere on your file system.
The type
command tells you the category of a command.
type echo # shell buitin
help echo
type ll # alias (typically defined in most linuxes)
type cp # executable
cp --help
declare -F # list all functions
The vast majority of commands that you are going to type are programs (scripts or binary executables). The list of directories containing programs is stored in the environment variable PATH
:
echo $PATH
Note that directory names are separated by the character :
.
If you want to add a directory, say /opt/bin
, to the PATH:
export PATH="/opt/bin:$PATH"
From now on, the shell will search for programs in /opt/bin
before scanning the other directories listed inPATH`. The program executed is the first encountered in the list. You can scan the list with:
which -a command
If you “screw up” the PATH, you will no longer have access to programs. In this situation, the best is to close the shell (by pressing Ctrl-D
) and open a new one. You can test this situation typing just:
PATH=
If you want a modification of the PATH
variable to be permanent, i.e. to be active each time you start a shell, add the export PATH=...
line to the file ~/.profile
.
If you happen to often type the same series of commands, it is a good idea to create a script, that is, basically, a text file gathering the sequence of commands to be executed. Then, you will just have to type the filename of this script to execute all the commands.
If it does not exist yet, create a bin
directory in your home folder:
mkdir $HOME/bin.
Use a text editor to create a file myscript
in this directory, and enter the following on the first line:
#! /bin/bash
Then type the series of commands (one per line) you want to be executed.
Save the file myscript
and enter the commands:
chmod +x ~/bin/myscript
PATH="$HOME/bin:$PATH"
You can now type myscript
on the command line to execute the series of commands.
To go further, you should learn how to use arguments to scripts.
Note that you write scripts in other languages than bash, e.g. python.
~/.bash_profile
, ~/.profile
, .bashrc
are scripts that are executed automatically when you start a shell. This allows you to set up your environment (e.g. the PATH, the Prompt, create aliases for common operations, ...)
There are four types of shells:
- login shells and non-login shells.
- interactive and non-interactive shells
If you connect to a remote computer with ssh remote
, you get an interactive login shell.
If you execute a command on a remote computer with ssh remote command
, the script or the command is executed in a login non-interactive shell.
If you are already log in, that is, you have open a session, and open a new terminal, you get an interactive non-login shell.
If you execute a bash script, it is launched in a non-interactive, non-login shell.
Login shells execute ~/.profile
and ~/.bash_profile
.
Non-login shells only execute ~/.bashrc
, not ~/.profile
nor ~/.bash_profile
Anything that should be available to graphical applications OR to sh
(or bash
invoked as sh
) MUST be in ~/.profile
, not .bashrc
(If you launch a graphical application not from the terminal, it only knows about the environment that was created at login. In particular, it will not know about stuff in .bashrc
) .
~/.profile
has the stuff NOT specifically related to bash, suchas environment variables (PATH and friends)
- Anything that should be available only to login shells should go in
~/.profile
~/.bashrc
has anything you’d want at an interactive command line (Command prompt, EDITOR variable, bash aliases)~/.bashrc
must not print anything on the terminal. This could screw up sftp for example.~/.bash_profile
should just load .profile and .bashrc (in that order)- Make sure that
~/.bash_login
does not exist.
See:
- https://superuser.com/questions/789448/choosing-between-bashrc-profile-bash-profile-etc
- https://stackoverflow.com/questions/902946/about-bash-profile-bashrc-and-where-should-alias-be-written-in
- http://mywiki.wooledge.org/DotFiles
You can use ln -s to create shortcuts to folders that you quickly need to access.
ln -s myproject_buried_in_some_very_deep_path_too_long_to_type ~/myproject
After that, you will just need to type:
cd ~/myproject
If you are tired of typing intermediate directory names when changing directory, check out the Directory Bookmarks functions for bash described in this linux journal article about dirb.
Download https://raw.githubusercontent.com/icyfork/dirb/master/dirb.sh in your $HOME
folder and add the following line to the file $HOME/.bashrc
:
source $HOME/dirb.sh
Once installed, you can save bookmarks for specific directories (command s
) and later jump into them directly (command g
). Here are all the available operations:
s Save a directory bookmark
g go to a bookmark or named directory
p push a bookmark/directory onto the dir stack
r remove saved bookmark
d display bookmarked directory path
sl print the list of directory bookmarks
sl -l long list
sl -p path list
- Another possibility is to use cdhist, a tool that replaces the
cd
command by a new version that, when you type
cd --
, list the recently visited directories and let you select one. It is a python script, therefoere, to install it, you need to type:pipx install cdhist
And add the following lines to ~/.bashrc
:
if type cdhist &>/dev/null; then
. <(cdhist -i)
fi
Install the fuzzy file finder fzf
sudo apt install fzf
Then, the command fzf
will let you explore the files in the working directory interactively
Later, we describe tools to find files (find, fdfind, grep, ack, ag).
xdg-open file
It may happen that a program monopolizes most of the CPU, but does not longer respond to input. Such a program is crashed and should be “killed”.
For applications running in a terminal, first try to press Ctrl-C
.
If this does not work, or if the application is running in its own window but refusing to close, open a terminal and type:
pkill program_name
You can also use the command ps -ef
to locate the application and note down the “process identification number” in the PID
column. Then, type:
kill PID
(in place of PID, use the number associated to the process listed in ‘ps’ output). Check if the program was destroyed with the ps
command; if not:
kill -9 PID
If the whole graphics system no longer responds, you can try to open a text mode terminal with Ctrl-Alt-F1
or Ctrl-Alt-F4
, log in and kill the programs that causes problem. Sometimes, the only solution is to kill Xorg
, the display server).
It the keyboard does not repond anymore, before switching off the computer, you can try to connect from another computer on the same network using ssh
and to kill the applications or do a proper shutdown (typing ‘halt’ on the command line).
Powering off is easy, just type:
sudo shutdown
You my want to specify a delay:
sudo shutdown --halt +1 # one minute delay
You can cancel the shutdown during the delay with:
sudo shutdown -c
If you want to reboot the system:
sudo shutdown -r now
If your workstation is switched off, but you can log to another linux computer on the same local area network, you might be able to power it on if you have authorized Wake on lan (WOL) in your station's BIOS parameters.
First, you need to know the MAC address of your computer's network interface (using ip a
when the computer was on).
Say the MAC address is "c8:f7:50:bc:ea:f5", then the command:
wakeonlan c8:f7:50:bc:ea:f5
launched on the terminal of another computer will power on your computer.
See http://doc.ubuntu-fr.org/wakeonlan
To get a list of available printers:
lpstat -p -d
To check the status of all printers:
lpstat -a
To print file.pdf
(or more precisely to put in the printing queue) of the printer printername
:
lpr -P printername file.pdf
To print two copies of a file
lpr -# 2 filename.pdf
To print 2 pages per side:
lpr -o number-up=2 -o sides=two-sides-long-edge filename.pdf
To remove a printing job:
lprm job-id
(job-id
is the number reported by the lpr
or lpstat
commands).
If you use the same printer most of the time, you can create a script like the following in your ~/bin
directory:
#! /bin/sh
export PRINTER=my-beautiful-printer
lpr -P "$PRINTER" -o media=A4 "$*"
In case of printing problem, first Check that that the cups service is running:
systemctl status cups.service
If you need to manage or add printers, open a browser on http://localhost:631
Check out Linux 101: Manage printers and printing for more information.
To use a one-time password:
To encrypt file.txt
:
gpg --symmetric file.txt # this will create file.txt.gpg
rm file.txt # do not forget to remove the unencrypted file
To decrypt it:
gpg -o file.txt --decrypt file.txt.gpg
Note that it is also possible to use gpg to generate a private/public key pair to sign documents (see https://tutonics.com/2012/11/gpg-encryption-guide-part-1.html )
Use the programs xranrd
and arandr
arandr
xrandr --output eDP1 --rotate left
If you have a nvidia graphics card, you can also use nvidia-settings
A secure method to connect to a remote computer:
ssh computername
or, if your login id on the remote computer is different than the one on the local computer.
ssh login@computername
If you plan to launch graphical application on the remote computer, you need to add the -X
option:
ssh -X login@computername
Note: you may need to run xhost +
on the local (client) computer.
If you often connect to a computer, you can create an entry in $HOME/.ssh/config
:
Host myserver
Hostname gozilla.example.com
User mickey
Then you will have just to type ssh myserver
to log in.
To have TAB completion on server names contained in .ssh/config
, create a file /etc/bash_completion.d/ssh
with the following content:
_ssh()
{
local cur prev opts
COMPREPLY=()
cur="${COMP_WORDS[COMP_CWORD]}"
prev="${COMP_WORDS[COMP_CWORD-1]}"
opts=$(grep '^Host' ~/.ssh/config ~/.ssh/config.d/* 2>/dev/null | grep -v '[?*]' | cut -d ' ' -f 2-)
COMPREPLY=( $(compgen -W "$opts" -- ${cur}) )
return 0
}
complete -F _ssh ssh
Note that:
- the client computer must have the ssh client (
sudo apt install openssh-client
) - the remote computer must be running a
sshd
server (runsudo apt install openssh-server
on it).
You can troubleshoot connection issues with
ssh -vv login@computer
To avoid having to type your login password each time you use ssh or scp, you can setup SSH to use public and private keys to perform the authentification automagically.
First, you must generate keyfiles, once, on your local computer. To do so:
ssh-keygen
This generates, among other files, a public key stored in a file ~/.ssh/id_rsa.pub
). You now need to copy this key in the file ~/.ssh/authorized_keys
on the remote computer you want to connect to. This can be done with:
ssh-copy-id login@remotecomputer
If you have left the passphrase empty, you can know use ssh
or scp
without entering your password. But so can do anyone who has access to your account on the local computer.
So you may prefer to use a passphrase. To avoid having to type it each time you log to the remote computer, copy the following lines in your ~/.bash_profile
:
eval `ssh-agent`
ssh-add < /dev/null
You will be prompted for the passphrase only once: when you login on the local computer (See the explanations about ssh-agent
at http://mah.everybody.org/docs/ssh).
ssh login@computername command
Beware the ~/.bashrc
script on the remote computer will not be executed because ssh launches a non-interactive, non-login shell. Thus the remote PATH
may not be what you expect! (solution: set the PATH
in .profile
, not .bashrc
)
Once connected on the remote computer, execute:
tmux
When you want to leave, press Ctrl-b d
. The terminal is detached but not closed.
Next time you connect to this remote computer, to continue your work, you can access the session:
tmux a
See https://danielmiessler.com/study/tmux/ for a primer on tmux, or read the book Tmux 2: Productive Mouse-Free Development by Brian Hogan.
scp -r localdir remotelogin@remotecomputer:remotedir
rsync -avh localdir/ remotelogin@remotecomputer:remotedir
tar -cf - dir | ssh login@remotehost tar -xvf -
sshfs login@remotecomputer:path local_path
To allow graphical applications running on the server to display their windows on the local computer, when using ssh:
X11 forwarding needs to be enabled on both the client side and the server side.
On the client side, the -X (capital X) option to ssh enables X11 forwarding, and you can make this the default (for all connections or for a specific conection) with ForwardX11 yes
in ~/.ssh/config
.
On the server side, X11Forwarding yes
must be specified in /etc/ssh/sshd_config
. Note that the default is no forwarding (some distributions turn it on in their default /etc/ssh/sshd_config
), and that the user cannot override this setting.
The xauth
program must be installed on the server side. If there are any X11 programs there, it’s very likely that xauth
will be there. In the unlikely case xauth
was installed in a nonstandard location, it can be called through ~/.ssh/rc
(on the server!).
Note that you do not need to set any environment variables on the server. DISPLAY
and XAUTHORITY
will automatically be set to their proper values. If you run ssh
and DISPLAY
is not set, it means ssh
is not forwarding the X11 connection.
To confirm that ssh
is forwarding X11, check for a line containing Requesting X11 forwarding in the ssh -v -X output
. Note that the server won’t reply either way, a security precaution of hiding details from potential attackers.
You cand launch commands in background by appending a &
sign at the end of the command line:
gedit &
(Note if a command is already running, blocking the terminal, you use Ctrl-Z
and then the built-in command bg
to put the process in background)
You can use this to launch several processes in parallel.
Consider the following shell script test.sh
that just print its arguments and sleeps for 5s:
#! /usr/bin/bash
echo "Args: $@"
sleep 5s
echo "Done"
Set the executre flag:
chmod +x test.sh
You can launch 5 instances of it with a for loop
:
for i in $( seq 1 5 ); do (./test.sh $i &); done
If you want to , to avoid mangling the standard output, you can save the output of each process in a different file:
for i in $( seq 1 5 ); do (./test.sh $i >test_${i}.out.txt &); done
To run each command in a different terminal (you may need to install xterm
):
for i in $( seq 1 5); do (xterm -hold -title "Process $i" -e ./test.sh $i &); done
The xargs
and parallel
commands also allows one to launch the same command in parallel with different parameters:
seq 1 5 | xargs -P 5 -n 1 ./test.sh
parallel ./test.sh ::: $( seq 1 5 )
To know more, read their documentations...
To display the network node name (also called the hostname
):
hostname
or
uname -n
To know your public address on the Internet:
sudo apt install curl
curl ifconfig.me
To know your IP address on the local area network:
ip addr
(you must identify the physical interface (ethernet card or wifi card) and check for the inet
line)
df -hT -x squashfs -x tmpfs
I actually added the following in my .bashrc
:
alias df="df -hT -x squashfs -x tmpfs"
if you need to make space you can search for large folders or files using:
ncdu
du -h | sort -hr | less
If there is a quota system that limits the amount of space you can use on your account, you can check how much is available:
quota -s
lsblk | grep -v loop # excludes loop devices
blkid
To list the processes currently running:
ps auf
ps axuf # also show process no tied to a terminal
The most important columns are TIME
and RSS
which show the time used by process since it started and the amount of real memory it uses.
If you want to list just some programs, for example matlab
, type
pgrep -a matlab
For a real-time display of processes, you can use top
or htop
but a more comprehensive too is glances
:
glances
Not only does it display CPU and memory usage, but also DISK I/O and network I/O. You can sort processes, for example, by CPU usage, etc (Press h
in glances to see the help). Glances is extremely useful to identify bottlenecks (see https://livebook.manning.com/book/linux-in-action/chapter-13/74)
You may have to install it with pip install glances
or sudo apt install glances
.
Note: there exist other process monitoring tools: atop, btop, htop, nmon
Sometimes, it can useful to find the process that owns an open file:
lsof filename
(See http://www.thegeekstuff.com/2012/08/lsof-command-examples/)
neofetch
sudo inxi -b
nvidia-smi # if you have nvidia GPUs
To check how many CPU/cores are available on your machine:
lscpu -e
lscpu
To check the total amount of RAM installed on your computer and how much is currently being used by Linux:
free -h
Which Linux distribution is running:
inxi -b
lsb_release -a
Note: you may need to install the packages inxi
and lsb-core
:
# deb based linuxes: sudo apt install lsb-core
# rpm-based linuxes: yum install redhat-lsb-core
# redhat/fedora: dnf install redhat-lsb-core
Which version of the linux kernel is running:
uname -a
Note: for improved real-time performance, you need to see PREEMPT_DYNAMIC
. You can also check the value of CONFIG_HZ
with
grep ^CONFIG_HZ /boot/config-`uname -r`
If CONFIG_HZ is less than 1000, it may be worth installing a realtime kernel to reduce latencies and the probability of audio dropouts:
sudo apt install linux-lowlatency-hwe-22.04
lshw -short
hwinfo --short
lspci
inxi -b
sudo apt install lm-sensors hddtemp
sudo sensors-detect
sensors
You can then install psensor
to have a GUI monitoring the temperatures:
sudo apt install psensor
psensor
You can monitor your system with glances
:
glances -t 5
or with htop
:
htop -d 50 --sort-key PERCENT_CPU
htop -d 50 --sort-key M_RESIDENT
There are more specialized tools that focus on subsystems. For example, you can monitor the global activity of the CPUs with:
mpstat 5
To monitor the memory usage in real-time:
vmstat -S M 10
If any of the indicators si
(swap in
) or so
(swap out
) are high, your computer lacks memory and is using the swap (memory on disk).
You can check the file input/ouput volume and speed on the local drives:
iostat -x 2 5
iostat -h -d 10
Check the speed of your ethernet connection. Three tools are available:
mii-tool
ethtool
iperf
Or the general network performance:
netstat -i 10
Large TX-ERR
or RX-ERR
indicate a problem.
sudo netstat -tulpn
systemctl -l -t service | less
You can simply use hdparm and `dd`:
sudo hdparm -tv /dev/sdc1 # read test
dd if=/dev/zero of=/disk/temp oflag=direct bs=128k count=1G # write test
(See https://linuxconfig.org/how-to-benchmark-disk-performance-on-linux)
For a more detailed analysis, install and run `fio`:
man fio
fio --name TEST --eta-newline=5s --filename=fio-tempfile.dat --rw=read --size=500m --io_size=10g --blocksize=1024k --ioengine=libaio --fsync=10000 --iodepth=32 --direct=1 --numjobs=1 --runtime=60 --group_reporting
fio --name TEST --eta-newline=5s --filename=fio-tempfile.dat --rw=write --size=500m --io_size=10g --blocksize=1024k --ioengine=libaio --fsync=10000 --iodepth=32 --direct=1 --numjobs=1 --runtime=60 --group_reporting
(from https://askubuntu.com/questions/87035/how-to-check-hard-disk-performance)
glmark2
sudo mkdir -p /mnt/ramdisk
sudo mount -t tmpfs tmpfs /mnt/ramdisk -o size=1024M
sudo chown `whoami`:`whoami` /mnt/ramdisk
ls -al /mnt/ramdisk
Two tools can be used to monitor power usage:
sudo powertop
powerstat
If you have a nvidia card:
nvidia-smi
ss -tr
sudo apt-get install -y lynis rkhunter clamav clamav-daemon -y
sudo lynis audit system
sudo rkhunter -c
systemctl start fwupd
# list devices that support firmware updates
fwupdmgr get-devices
# updating
fwupdmgr refresh
fwupdmgr get-updates
fwupdmgr update
As far a the computer is concerned, the identity of the current user (its user_id), can be printed with:
whoami
Note that your login name and home directory are stored in the environment variables LOGNAME
and HOME
.
Each login is associated to a UserID (UID), an integer, and to a list of GroupIDs (GUID). You can list the information associate to the current login:
id
To see who is currently logged on the system, use
who
or more simply:
w
If you are superuser, you can see a journal of the logins with the command:
sudo last
To determine a person behind an user_id, use finger
:
finger <user_id>
To temporally become newuser
:
su - newuser
Of course, you will be prompted for newuser’s password.
If you want to become root
:
sudo -i
When you are done, type:
exit
To change your password on the local system:
passwd
pwgen 10 --symbols
generates a 10 character long password with at least one special character
Run it online at https://pwgen.io/en/
To change your login shell, e.g. from /bin/csh
to /bin/bash
:
chsh -s /bin/bash
Check which groups you belong to using id
, then use
newgrp group
From now, the files and directories you create will belong to group group
To modify the group of already existing files in directory dir
:
chgrp -R group dir
Each login is associated to a number called the UID
. If for any reason you need to change your UID number, here is how to do it:
usermod -u <NEWUID> <LOGIN>
groupmod -g <NEWGID> <GROUP>
find / -user <OLDUID> -exec chown -h <NEWUID> {} \;
find / -group <OLDGID> -exec chgrp -h <NEWGID> {} \;
usermod -g <NEWGID> <LOGIN>
sudo usermod -aG sudo userlogin
Of course, you need to be in the list of sudoers yourself to be able to execute this command.
Using sudo
is better than using su
, check out why at https://phoenixnap.com/kb/sudo-vs-su-differences
To know the current working directory:
pwd
To change the current working directory:
cd subdirectory # move down inside a subdirectory
cd .. # move up in the hierarchy of directories
Note that you can always go back to your home directory by just typing cd
(without argument).
ls # list (non hidden) files and subdirectories in the current working directory
ls -A # list all files (including hidden ones)
ls -1 # list in a single column
ls -l # show detailed information (filesize, modification date,...)
ls -t # sort by modification date (most recent first)
ls -Slt # sort by size (largest first)
ls PATTERN
where PATTERN is a globbing pattern which can contain “wildcards” characters such as *
or ?
:
*
| matches any string | |?
| matches any character | |my*
| filename starting with my
| |my?
| filename of 3 characters stating with my
| |*xyz*
| filename containing xyz
| |*.tar*
| filenames finishing with .tar
| |*{md,txt}
| filenames ending in either md
or txt
| |*.[ch]*
| filename ending in .c
or `.h** | |By default, ls
only lists the files in the current working directory. To recursively visit the subdirectories:
ls -R
ls **/**.py
**
will match the first-level subdirectories. With the option shopt -s globstar
, subdirectories at all levels are visited.
To only display subdirectories:
ls -d */ # only directories
tree -d # Recursively
tree -d -L 2 # limit depth to 2
To copy a file inside the same directory, giving it name2:
cp file1 file2
To copy a file from the current directory to the existing directory target_dir
:
cp file1 target_dir
To copy all the files from the current directory to another directory:
cp * target_dir
To do the same thing but showing a progress bar:
rsync --info=progress2 * target_dir
To rename a file:
mv file1 file2
To move a file to the existing directory dir
:
mv file1 dir
To delete a file:
rm file
To avoid being asked for confirmation:
rm -f file
To create a new directory:
mkdir -p newdir
To copy the directory dir
inside the destination directory destdir
:
cp -a dir destdir
(Note: the -a
option does a recursive copy, that is, includes the subdirectories and preserves the attributes of files)
Alternatively, you can use rsync
:
rsync -a --info=progress2 dir/ destdir
To move the whole directory dir
inside the existing destdir
:
mv dir1 destdir
To rename directory dir
as dir2
:
mv dir dir2
To delete the directory dir
and all its content:
rm -rf dir
Here is a script that replaces filenames by creation date (this can be useful for a photo album)
#! /bin/bash
for fullfile in "$@";
do
filename=$(basename "$fullfile")
extension="${filename##*.}"
filename="${filename%.*}"
mv -n "$fullfile" "$(date -r "$fullfile" +"%Y%m%d_%H%M%S").${extension}";
done
When you use ls -l
to list the files in a directory, the first string of characters, made of x
, r
, w
, -
… specifies the access rights (Consult Understanding file permissions on Unix: a brief tutorial)
To allow everybody to read a file aga
in the current directory:
chmod a+r aga
To allow everyone to enter a directory mydir
and read its content:
chmod a+rx mydir
To make all subdirs and files readable by everyobody:
find -type d -exec chmod a+rx '{}' '+'
find -type f -exec chmod a+r '{}' '+'
If, when using ls -l
, there is a +
sign is trailing the rights, it means that ACL (Access Control List), is set on the files or directories. The chmod command will not work: you must then use the getfacl
and setfacl
commands to list or modify the access/write rigths
To avoid copying a file in several places on the same disk, it is a better idea to use a hard link:
ln existingname newname
Thus the same file can have several names (and be in several directories at the same time). Importantly, this only works if the directories are on the same filesystem.
To create a symbolic link (somewhat similar to a ‘shortcut’ in Windows):
ln -s filename newname
If you delete or move the file, the symbolic links will be ‘dangling’.
To find and remove dangling links in a directory:
symlinks -rd directory
The classic unix command to find files is, well, find
.
We describe it below, but we first introduce a simpler and user-friendly alternatives: fd
.
Examples of usage:
fd statement # search for files/directories containing the string "statement" in their name
fd -t f statement # restrict the search to files (not directories)
fd pdf ~/Downloads/ --changed-within 1hour
You can search for filenames matching a regular expression:
fd 'April.*docx$'
Features of fdfind:
- Regular expression (default) and glob-based patterns
- Very fast due to parallelized directory traversal
- Uses colors to highlight different file types (same as ls)
- Supports parallel command executio
- Smart case: the search is case-insensitive by default. It switches to
- case-sensitive if the pattern contains an uppercase character*.
- Ignores hidden directories and files, by default.
- Ignores patterns from your .gitignore, by default.
Note
You may need to install fd
using sudo apt install fd-find
or from https://github.com/sharkdp/fd, and define alias fd=fdfind
.
Another must know user-friendly search tool is ag
which allows to spot text files containing a given string or regular expression:
ag --python "import numpy" # search python files that import numpy
Note
To install ag
under Ubuntu: sudo apt install silversearcher-ag
.
find
is the classic command, which is complex but powerful. THe basic syntax is:
find -name pattern
where pattern
can be a string, or a glob pattern (not a regular expression):
find -iname 'filename.txt'
find -iname '*.doc'
The last command will list all *.doc
files in the current directory and its subdirectories. The depth of subdirectories to visit can be limited:
find -maxdepth 2 -name '*.doc'
If you prefer regular expressions to glob patterns, use the option -regex
instead of -name
:
find -regex '.*.txt'
With -o
you can specify an ‘or’. For example, to search for for files with extension nii
or img
:
find \( -name '*.nii' -o -name '*.img' \) # files ending in .nii or .img
With !
, you can negate a search:
find ! -name '*.nii' # all files except those ending in .nii
You can specify a time-range:
find -mtime 0 # find the files created or modified in the last 24hours
find -mtime +30 -mtime -60 # find files modified in the last 30-60 days
find -newermt 20171101 ! -newermt 20171201 -name '*.pdf' -ls # find pdf files modified between two dates
You can specify that you only search for, e.g., directories, using the -type
argument:
find -type d # list all subdirectorectries
find -type d -mtime -10 # find the directories created or modified in the last 10 days:
You can find and delete all empty directories:
find . -type d -empty -print
find . -type d -empty -delete
You can filter on permissions
find -perm -o+x -ls -type f # list all file with the execute flag set on 'others'
You can also execute a command on each file:
find -name '*~' -exec rm '{}' '+' # delete all files '*~'
find -name '*.py' -exec mv -t path '{}' '+' # move all py files to path
find -name '*.txt' -print0 | xargs -0 grep -l Alice # show files
Note that xargs
can be parallelized with the -P option:
find -name '*.nii' -o '*.img' -print0 | xargs -0 -P 10 gzip # gzip all image files
Consult info find
and info xargs
for more information.
To accelerate file search, you can generate a database of all filenames on your filesystem. First of all, make sure you have installed plocate
:
sudo apt install plocate
sudo updatedb
Enable an automatic update of the database:
sudo systemctl enable plocate-updatedb.timer
sudo systemctl start plocate-updatedb.timer
And then use the command:
plocate PATTERN
Note that the plocate will return all files where PATTERN matches any substring in the full pathname (including directories).
Read the manual:
man plocate
grep PATTERN file
where PATTERN is a regular expression (See man grep
).
To search files recursively in subdirectories, you can combine find
and grep
:
find -type f -name "*.tex" -print0 | xargs -0 grep -n PATTERN
But this is complex! An interesting alternative is to use ack
(https://beyondgrep.com/). By default, it does a recursive search and it can focus on certain file types.
ack --python -w TOKEN # search only python file matching on word 'TOKEN'
To install ack
under ubundu:
sudo apt install ack-grep
Another search tool is ag
http://conqueringthecommandline.com/book/ack_ag:
sudo apt install silversearcher-ag
Tools like grep
, ack
andag
are useful to search within text files but pretty useless for binary files. If you need to search within .pdf
or .doc
files, you first need to extract the textual content and then index it. Then, you will be able to search files by their content. To this end, you can install and use a tool like recoll
(see http://www.lesbonscomptes.com/recoll/). One issue though it that the index can quickly grow very large.
To list all the lines that differ between file1 and file2:
diff file1 file2
meld
provides a nicer, graphical way to show the differences between two files or two directories.
meld file1 file2
When comparing text file, you may want to ignore changes in whitespaces (e.g. wrapping of paragraphs), then use wdiff
.
wdiff file1.txt file.txt
To compare two latexfiles:
latexdiff file1.tex file2.tex
To create a patch listing the changes from version1 to version2:
diff -aur version1 version2 >dir2.diff
To apply the patch to version1 and generate version2:
patch -p1 <dir2.diff
To compare two directories:
diff -r --brief dir1 dir2
diff
compares the contents of the files. For large directory, this may be too slow. To run a faster comparison based on file sizes, you can use:
rsync --dry-run --recursive --size-only -i source/ target/
unison
To back up my laptop, I use rsnapshot. I use an external harddrive with a large ext4 partition (~4 times the sizae of my laptop harddrive).
sudo apt install rsnaphost
Configuring rsnapshot essentially consists of editing /etc/rsnapshot.conf
to specify where to save snapshots. In my case:
snapshot_root /media/cp983411/WD_BLACK/rsnapshot/
Another nice backup utility, with a graphical interface, is:
backintime
It can be set up to automatically start so that you just have to plug your backup hardrive to performe a backup. Check out http://backintime.readthedocs.io.
wget URL
wget --recursive --level 2 --no-cookies --page-requisites --convert-links URL
curl address
To send folder
to a remote host:
rsync -azv folder username@hostname:path
To reverse the direction of transfer, simply swith the two arguments.
An interesting option is --delete
which makes the remote a mirror of the local.
See https://tutonics.com/2012/05/netcat-basics.html
Copy remote folder locally:
scp -r username@hostname:path_to_folder .
Send local folder to remote host:
scp -r folder username@hostname:path
If you need to transfer files using the ftp protocol, you can use the following clients
ncftp
lftp
Transfer fil
Another approach to synchronise dirs is to use git repositories.
Learn about git by reading https://git-scm.com/book/en/v2
See also git-annex
git push --mirror git@github.com:username/project.git
killall syndaemon
syndaemon -i 1 -KRd
sudo rmmod psmouse
sudo modprobe psmouse
Try Ctrl-Alt-F1
to open a terminal. From there, you might be able to do:
sudo shutdown now
Alternatively, press Alt+PrintScr
, and, keeping this key pressed, type, slowly, reisub
. This mysterious sequence is explained at https://linuxconfig.org/how-to-enable-all-sysrq-functions-on-linux#h6-the-sysrq-magic-key or https://en.wikipedia.org/wiki/Magic_SysRq_key
sudo brightlight -r # read
sudo brightlight -i 10 # increase
sudo brightlight -d 10 # decrease
or
xbacklight -set 50
or
xrandr --output eDP1 --brightness 0.5
Assuming that xscreensaver
is running in the background.
xscreensaver-command -lock
or:
i3lock -d 30 # if you use i3wm
systemctl suspend
systemctl hibernate
Note: To hibernate on disk, the size of the swap partition must be larger than the RAM size.
systemctl reboot
systemctl poweroff
Make sure to have ImageMagick installed (e.g. sudo apt install imagemagick
on a Debian-based system)
To get information about an image:
identify image.png
To display an image (gif, .jpg, .png, .tiff, eps, …) use:
display file.gif
eog image.png
To convert from one format to another:
convert file.jpg file.png
To resize an image:
convert img.png -resize 66% img_small.png
convert img.png -resize 400x400 img_400.png
To juxtapose several images:
montage -tile 4x4 *.png -geometry 1024x768 output.png
To superimpose images:
composite img1.png img2.png result.png
For more complex manipulations of bitmap image, I mostly use The Gimp
gimp file.jpg
To manipulate photographs, checkout:
To draw on canvas (with pencils, brush, …)
To edit vector graphics files, e.g. .svg
:
inkscape
To create graphs:
dot
To plot data, I use R
or `Python: :: import matplotlib.pyplot as plt import numpy as np Take a screenshot ~~~~~~~~~~~~~~~~~ To take a snapshot, that is, copy a portion of the screen into an image file, you can use ImageMagick’s command
import: :: import file.png You will then be able to select a rectangle on the screen with the mouse, which will be copied in
file.png. Other screenshot programs include
gnome-screenshot,
ksnapshot,
scrot,
maim\ … See https://wiki.archlinux.org/index.php/Screen_capture for a list. Make a screencast ----------------- Voir http://www.linuxlinks.com/article/20090720142023520/Screencasting.html Under i3, see https://github.com/synaptiko/.files/blob/4a6a549dfe0c22d19f38e32129b5c05de2bb6d34/i3/record-screen.sh Sound ----- Assuming that your Linux distribution is running the pulseaudio sound server — which can be checked with
pactl list—, install
pavucontrolto control the sound levels and which sound card each software is using. Connect a MIDI instrument ~~~~~~~~~~~~~~~~~~~~~~~~~ Follow the instructions at http://tedfelix.com/linux/linux-midi.html. In a nutshell: :: sudo apt install jackd2 jack-tools fluidsynth aconnectgui vmpk qjackctl qsynth fluid-soundfont-gm 1. To avoid potential latencies, you may want to install a kernel with the PREEMPT option: sudo apt-get install linux-lowlatency-hwe-20.04 2. Launch
qjackctl, in the setup tab, set Frame/period to 128 to reduce latency, ans press ‘start’ 3. Use aconnectgui to connect your MIDI keyboard 4. Launch qsynth, add the soundfounds in setup and restart it. 5. In qjackctl, use connect and the patchbay. Miscellaneous ------------- Access files on a data CD or on a floppy ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ With some Linux systems, you just insert the CD or the floppy and the content become available in the directory
/mnt/cdromor
/mnt/floppy: :: ls /mnt/cdrom ls /mnt/floppy If the floppy is not write-protected, you can create or copy files in /mnt/floppy just like in any ordinary folder. Note that if you have several cdrom or floppy drives, they may have names cdrom1, cdrom2, floppy1,… In some Linux systems, it is necessary to manually *mount* the cdrom or the floppy before accessing the files, and *umount* it before ejecting it. For the cdrom: :: mount /mnt/cdrom ls /mnt/cdrom ... umount /mnt/cdrom eject For the floppy: :: mount /mnt/floppy ls /mnt/floppy umount /mnt/floppy If you get an error message like
mount: only root can do that, ask the system administrator to grant you right to mount floppies by adding the
useroption the configuration file
/etc/fstab. More information in the manual pages of
mountand
fstab: :: man mount man fstab Concerning floppies, some systems have
mtoolsinstalled (see
man mtools') which provide the\ mdir\
and\ mcopy\
commands that emulate the old DOS commands\ dir\
and\ copy`. It is not necessary to mount the floppy to use them. Format a floppy ~~~~~~~~~~~~~~~ To format the floppy with an ext2 filesystem, and mount it: :: fdformat /dev/fd0 mkfs -t ext2 /dev/fd0 mount -t ext2 /dev/fd0 /mnt/floppy This floppy can be read only on other linux systems. To be able to read it under Windows/DOS, you should use a DOS filesystem with mkdosfs in place of mkfs -t ext2: :: mkdosfs /dev/fd0 Split a large file on several floppies ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ First compress the file, with gzip or bzip2 (see section\ `41 <#compress>`__). If it still does not fit on a single floppy (1.4Mb), you can use the command split: :: split -b1m file This create a series of x?? files which you can copy on separate floppies. To reassemble the files: :: cat x* >file Rip an audio CD ~~~~~~~~~~~~~~~ to extract all tracks from an audio CD: :: cdparanoia -B To just extract one track: :: cdparanoia -w track_number file.wav If you prefer GUI, you can open
konqueror, and type ‘
audiocd:/’ in the address bar. This will show you the content of the CD, which you can copy somewhere else. Copying from the mp3 or ogg folders will do the automatic translations for you. There are various programs with graphical interface which allow you to rip audio CD:
gripand
kaudiocreator,
rhythmbox. Convert from wav to mp3 ~~~~~~~~~~~~~~~~~~~~~~~ I use `lame <http://lame.sourceforge.net/>`__: :: lame file.wav file.mp3 Convert from wav to ogg vorbis ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I use `oggenc <http://www.gnu.org/directory/audio/ogg/OggEnc.html>`__: :: oggenc file.wav -o file.ogg Rip an Audio cd into mp3 or oggenc ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ You could write a script calling
cdparanoiathen
lamebut there is a nifti command line tool,
abcde, which queries music databases to find the tracks’ song titles. :: abcde -o mp3 # rip an audio cd track and converts into mp3 If you prefer a GUI, use `asunder <http://littlesvr.ca/asunder/>`__ Rip a DVD ~~~~~~~~~ Use (handbrake)[https://handbrake.fr/] Create a data CD ~~~~~~~~~~~~~~~~ 1. Gather all the files you want to save in a given directory, e.g.
/tmp/mycd2. Create an iso image: :: mkisofs -o cd.iso -J -R /tmp/mycd ls -l cd.iso Check that the resulting file
cd.isofile is not too large to fit on the CD; if it less than 650Mb, this should be ok. 3. Record on the cd (you must be root). You must know which is the device is associated to the CD writer drive. :: cdrecord -scanbus To determine the x,y,z scsi coordinates of your cd writer. If it does not appear listed, it may be because the ide-scsi parameter was no passed to the Linux kernel (See the HOWTO about CD Writing). To record, do: :: cdrecord dev=x,y,z -multi speed=0 -data cd.iso Create an audio CD ~~~~~~~~~~~~~~~~~~ To record on an audio CD all the
*.wavfiles which are in the current directory: :: cdrecord dev=x,y,z -pad speed=0 -audio *.wav (x,y,z must be replaced by the numbers returned by cdrecord -scanbus) Make backups ~~~~~~~~~~~~ You can write backup scripts using
rsyncbut it has already been done many time. I have used `backintime <https://backintime.readthedocs.io/en/latest/>`__, but `borgbackup <https://borgbackup.readthedocs.io/en/stable/>`__ looks interesting. Connect to a bluetooth device ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :: sudo service bluetooth start sudo service bluetooth status rfkill list rfkill unlock 0: bluetoothctl power on devices scan on pair XXXXXXX connect XXXXXX Convert doc or odt documents to pdf ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :: libreoffice --headless --convert-to pdf *.odt List the hosts in a NIS domain ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ If you are connected on a local network administrated by NIS (
yellow pages), you can display the list of other computers on the network: :: ypcat hosts Mounting a Samba Share ~~~~~~~~~~~~~~~~~~~~~~ Assuming you have a SAMBA server with IP 192.168.0.50 :: smbclient -L 192.168.0.50 sudo mount -t cifs //192.168.0.50/BACKUPS /mnt -o username=chrplr,file_mode=0777,dir_mode=0777 Which shell is running? ~~~~~~~~~~~~~~~~~~~~~~~ When you enter commands on the command line in a terminal, the text you type is interpreted by a program called the ‘shell’. There are different shells that speak different dialects. To determine the shell you are communicating with, type: :: echo $SHELL Note: this does not work well for subshells: :: bash echo $SHELL csh echo $SHELL exit exit Get help. Find manuals ~~~~~~~~~~~~~~~~~~~~~~ Many commands have associated
man pages. To read the man page associated, for example, to the command
cp: :: man cp Some commands also have manuals in the form of
info files: :: info gawk On many linux systems, there is additional documentation in the
/usr/share/docfolder. The HOWTOs can be especially helpful. To browse them, install
dwww: :: sudo apt install dwww sudo a2enmod cgi sudo systemctl restart apache2 sudo dwww-index++ Then: :: dwww Cut’n paste ~~~~~~~~~~~ Cutting & pasting under linux is not always straigtfoward. This is due to the fact that there are various systems of cut’n paste cohabitating. To copy text, the following works with most applications: - Click the left button and drag the cursor over the text to be copied. - Click on the middle button to paste. Note that this is very convenient: there no need to explicitly ‘copy’ the text. If you use the window manager ‘kde’, there is a useful applet called ‘klipper’ located on the panel. Klipper keeps copies of the most recent clipboard contents. If a cut’n paste operation does not work, you may open klipper, select the relevant line, and retry to paste. It usually works. If it does not work, then you can try the Cut/Copy/Paste functions from the applications’ menus. Sometimes, it is necessary to save the region as a file in the first application, and insert this file in the second application. set up tap to click in i3 ~~~~~~~~~~~~~~~~~~~~~~~~~ :: sudo mkdir -p /etc/X11/xorg.conf.d && sudo tee <<'EOF' /etc/X11/xorg.conf.d/90-touchpad.conf 1> /dev/null Section "InputClass" Identifier "touchpad" MatchIsTouchpad "on" Driver "libinput" Option "Tapping" "on" EndSection EOF Tip from https://cravencode.com/post/essentials/enable-tap-to-click-in-i3wm/ Mount a partition of a usb drive ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Insert the USB drive, use
lsblkor
dmesgto find partitions, then use
pmountor
udisksctl: :: lsblk pmount /dev/sdb1 udisksctl mount -b /dev/sdb1 Check an SD card ~~~~~~~~~~~~~~~~ :: sudo apt install f3 lsblk # to find out which DEVICE the card is associated to f3probe sudo ./f3probe --destructive --time-ops DEVICE Setup an ethernet card to access the internet ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ You need to know IP, MASK, GATEWAY, DNS, HOSTNAME and DOMAIN: :: ifconfig eth0 IP netmask MASK up route add -net default gw GATEWAY netmask 0.0.0.0 eth0 hostname HOSTNAME echo "domain DOMAIN" >/etc/resolv.conf echo "nameserver DNS" >>/etc/resolv.conf Changing/Editing network connection ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ :: nmtui # text mode nmcli # text mode unity-control-center Install new software ~~~~~~~~~~~~~~~~~~~~ If it come as a .tar.gz and contain a configure script :: tar xzf package.tar.gz cd package ./configure --prefix=$HOME & make & make install This install the software in your home directory. To install it for every user, you need to omit the prefix option and be root when calling
make install``.
If you are on a apt-based system (Debian, Ubuntu):
sudo apt install packagename
If you have the .deb file:
sudo dpkg -i file.deb
If you are on a rpm-based linux system, to install an rpm file:
rpm -i package.rpm
To check if the package is correctly installed:
rpm -V package
To remove it:
rpm -e package
To check if, say, ghostscript is installed:
rpm -q ghostscript
You can get the list of all installed packages:
rpm -qa
To run, some programs need to access functions in dynamic libraries. Dynamic libraries have the extension .so. They are located in /lib, /usr/lib, /usr/local/lib…
To list the libraries needed by a program:
ldd program
After adding new a new dynamic library, e.g. in /usr/local/lib, you must run, as superuser:
ldconfig -n /usr/local/lib
It is possible, as a user, to tell linux to search libraries in a particular place, using the LD_LIBRARY_PATH variable. For more information about how dynamic libraries are accessed, consult the manual of ld.so:
man ld.so
sudo apt install cmatrix
cmatrix
The following works for me.
- Use a window manager that allows you that launch applications pinned on some workspace and to have the workspaces accessible by a fixed keystroke. The tiling window manager i3wm fits the bill.
- use Emacs/Spacemacs or vim as an editor
- Use Linux rather than Windows
- use anaconda3 for Python
- use git for projects
e xtension | file type | application(s) |
---|---|---|
txt | text or ascii file | cat, less (view), vim, emacs (edit) |
Adobe PDF | evince, okular (view, annotate), pdfarranger | |
ps, eps | postscript | gv (view) pstops (rearrange) ps2pdf (convert) |
html, htm | web page | links, konqueror, mozilla (view) soffice (create) |
png, jpg, gif… | graphic files | display (view) import (snapshot) convert (convert) gimp (manipulate) |
doc, xls, ppt | Office document | soffice |
sxc, sxi, sxw | OpenOffice document | soffice |
tex | TeX and LaTeX documents | tex, latex, pdflatex (process) |
dvi | Dvi documents | xdvi (view) dvips, dvipdf (convert to ps or pdf) |
gz, Z, xz, bzip | Compressed file | gunzip, xz, unxz, zip, bunzip2, bzip2 |
tar | tar archive | tar tf (view) tar xf (extract) tar cf (create) |
tar.gz | compressed archive | tar xzf (extract) |
tar.bz2 | Compressed tar archive | tar xjf |
zip | zip archive | unzip -l (view) unzip (extract) zip (create) |