The core.autocrlf setting is to prevent any line-end CR/LF related errors.
Note that the scripts do not required to be executable as should only be run dot-sourced.
git config --global core.autocrlf input
git clone https://github.com/roysubs/custom_bash
cd custom_bash
. ./custom_loader.sh
Auto-configure common settings to be cross-platform for most Linux distros (CentOS/Ubuntu/Debian etc). Specific tools for WSL are included but that only load if WSL is detected. By sourcing a single script custom_loader.sh, this sets everything up and puts the .custom script into ~ which is then invoked by ~/.bashrc at shell startup.
Goal was to keep all modifications outside of main bash configuration files, so that the changes do not pollute main configuration files to create a hard to maintain environment. Removal is then as simple as removing the two lines added at the end of ~/.bashrc where it calls ~/.custom.
The goal is to setup a usable / comfortable environment with essential settings in ~/.vimrc, ~/.inputrc and useful tools without requiring the time it normally takes to setup a sensible environment. Run git clone to take a copy and then customise by adding other settings as required. For fixes / additional tools that would be useful in the main project would benefit from, email [email protected] and we can add into the main project.
Default setups: Installation of various small packages that are a useful baseline to have on any new system. Setup of colour syntax highlighting on all cat output (by using the bat tool). Functions to help with various system tasks. Some simple banner setups that are cosmetic but can be nice to have. Many aliases and functions to make for a more consistent cross-platform setup.
Note that sudo apt install curl git -y is a useful first step as new installations are often missing these.
Option 1: Download full git project, then run . custom_loader.sh:
You must change the core.autocrlf settings when cloning on another WSL system or else all scripts will be broken
git config --global core.autocrlf input
git clone https://github.com/roysubs/custom_bash
git clone https://git.io/Jt0f6 # git.io shortened url
This is useful as the curl xxx | bash method runs everything immediately without prompting. After cloning the repo, run the loader with: . custom_loader.sh (it is best to dotsource like this as it has no execute permissions after cloning, and to allow it to dotsource .custom into the currect session when running). There is no need to use sudo to run this as require elevated tasks will invoke sudo inside the script.
Option 2: Download only the required files:
The project only requires the following two files, so this can be done instead of cloning the project.
curl -s https://raw.githubusercontent.com/roysubs/custom_bash/master/custom_loader.sh > custom_loader.sh
curl -s https://raw.githubusercontent.com/roysubs/custom_bash/master/.custom > .custom
Then run . custom_loader.sh from the current directory.
Option 3: Install the custom tools immediately
curl https://raw.githubusercontent.com/roysubs/custom_bash/master/custom_loader.sh | bash
We can use git.io to shorten github utl's, but this does not work for curl xxx | bash, so using long form here. This method is least resilient, but works fine in most situations.
Note: It is not possible to invoke a script with switches via curl (Option 3 above), which is why downloading first then running is a better approach (could instead have the script look for flag files on disk, e.g. .custom_install_system_wide then delete that file after running custom_loader.sh from Github, or maybe a specific variable). ToDo: Optionally make all of the above load system-wide, i.e. create /etc/.custom and make changes to /etc/bashrc, /etc/vimrc, /etc/inputrc instead of ~. If doing this, must also clean up ~ to remove the details there.
custom_loader.sh configures the environment. This simply adds a single line to .bashrc to dotsource ~/.custom for all new shell instances (whether console or terminal inside a GUI environment as there is a distinction!). It also installs / updates some small core tools that are generically useful (vim, openssh, curl, wget, dpkg, net-tools, git, zip, p7zip, figlet, cowsay, fortune-mod etc), then adjust some generic settings for ~/.vimrc, and ~/.inputrc. and describes the correct way to update localisation settings. ~/.custom will then be dotsourced into the currently running session to be immediately available without a new login. To install the latest version on any WSL distro, use the above curl command from inside the WSL instance.
.custom is called from ~/.bashrc using two lines that ensure that the changes only apply to interactive shells (i.e. will load equally in either ssh login shells or terminal windows inside a Linux Gnome/KDE UI, and will not load during non-interactive shells such as when a script is invoked, as discussed here).
How to uninstall
This is very easily done as the only main change are the lines in ~/.bashrc to call ~/.custom so that the main profile scripts are modified as little as possible, so will not mess up any system, just delete the two lines in ~/.bashrc that call .custom to remove everything. custom_loader.sh is also a template for how to easily extend for other preferred settings; simply clone the project and adjust to add remove your own settings as required.
The following is the full syntax for all steps for an Ubuntu distro (Ubuntu partnered with Microsoft for the WSL project so their images are probably the most stable). Note that each distro is an independent VM (running on Hyper-V), but they are completely managed by the OS and so have almost instant start times. WSL VM folders (before making changes and installing software) are usually around 1 GB per instance. WSL1 vs WSL2, Docker with WSL2 Backend
Note: As of Wind 10 (2004 or later), WSL can now be setup with a single command:
wsl --install # Just installs WSL
wsl --install -d Ubuntu # Install WSL and a named distribution (-d = --distribution)
wsl --install -d Debian # Install WSL and a named distribution (-d = --distribution)
wsl --list --online # Show compact and friendly names for available distributions
Current wsl --install -d: Ubuntu, Debian, kali-linux, openSUSE-42, SLES-12, Ubuntu-16.04, Ubuntu-18.04, Ubuntu-20.04
# Deprecated: dism.exe /online /enable-feature /featurename:Microsoft-Windows-Subsystem-Linux /all /norestart
# Deprecated: dism.exe /online /enable-feature /featurename:VirtualMachinePlatform /all /norestart
# Deprecated: iwr -Uri https://aka.ms/wslubuntu2004 -OutFile Ubuntu.appx -UseBasicParsing
# Deprecated: Add-AppxPackage .\Ubuntu.appx
# Install Location: C:\Program Files\WindowsApps\CanonicalGroupLimited.Ubuntu20.04onWindows_2004.2020.812.0_x64__79rhkp1fndgsc
wsl --set-default-version 2
# Note that C:\Program Files\WindowsApps is quite security-restricted, cannot enter easily even as Admin.
# Also, this all bloats the system drive, so I normally move WSL distros to D:\WSL\<distro-name> (and can port distros to other systems in this way also):
md D:\WSL\Ubuntu
wsl -l -v
wsl --export Ubuntu D:\WSL\Ubuntu.tar
wsl --unregister Ubuntu # unregister the WSL image
wsl --list # verify the distribution has been removed.
wsl --import Ubuntu D:\WSL\Ubuntu D:\WSL\Ubuntu.tar
# Unfortunately, Ubuntu will now use root as the default user. To go back to your own account:
ubuntu config --default-user <yourname>
# where <yourname> is the username you defined during installation of that distro.
How to setup multiple instances of the same Linux distro in WSL (using --import and --export)
It is often useful to have multiple copies of a distro available. With the above exported distro this can be done easily:
md D:\WSL\Ubuntu1
md D:\WSL\Ubuntu2
wsl --import Ubuntu D:\WSL\Ubuntu1 D:\WSL\Ubuntu.tar
wsl --import Ubuntu D:\WSL\Ubuntu2 D:\WSL\Ubuntu.tar
Other Distributions Awesome-WSL information, distributions, and tools.
# Optionally, can install alpine, arch, or fedoraremix from the Chocolatey repository:
choco install wsl-alpine # Use Chocolatey to get Alpine (not offered by Microsoft)
choco install wsl-archlinux # Use Chocolatey to get Arch (not offered by Microsoft)
choco install wsl-fedoraremix # Use Chocolatey to get Fedora (not offered by Microsoft)
wsl-fedoraremix, wsl-alpine, wsl-archlinux
# CentOS 8 Stream installer https://github.com/mishamosher/CentOS-WSL/releases/tag/8-stream-20210603
After extracting the zip into a folder, inside it you will see two files: rootfs.tar.gz, CentOS.exe.
Run CentOS.exe in order to extract the vhd (Hyper-V virtual hard-disk), and register it on WSL.
Now, right-click CentOS.exe again, and run as Administrator, and this time it will start CentOS.
Note: You will be root after this and need to create a user and run as that:
# useradd <user>, then passwd <user>, if you can want a simple password, CentOS will complain, ignore that.
# usermod -a -G wheel <user> # CentOS/Fedora use the wheel group (would be usermod -a -G sudo <user> on Debian/Ubuntu as uses sudo group)
You must do the above before closing the root console or you will have no sudo.
# .\CentOS8-stream.exe config --default-user <user> to default to the new user.
To uninstall CentOS: .\CentOS.exe clean (from PowerShell).
If accidentally delete the CentOS folder, when unzipping it again, also should clean it again: .\CentOS.exe clean.
# CentOS 8 installer https://github.com/mishamosher/CentOS-WSL/releases/tag/8.4-2105
# CentOS 7 installer https://github.com/mishamosher/CentOS-WSL/releases/tag/7.9-2009
# CentOS 6 installer https://github.com/mishamosher/CentOS-WSL/releases/tag/6.10-1907
Working only on WSL2 with vsyscall=emulate. More info: microsoft/WSL#5465
Detailed instructions:
Install via wsl --import or running CentOS6.exe
Make sure you use WSL2: wsl --set-version <Distro> 2
Configure WSL2 to use vsyscall=emulate: microsoft/WSL#4694 (comment)
You have to restart the Windows service LxssManager for the step above to take effect.
# Must reboot before using a distro, but first set all defaults to use WSL v2.
# Fedora Remix for WSL
# Fedora Remix 34.5 for WSL with Windows Terminal Theme
LXRunOffline.exe install -n Fedora29 -f .\fedoraremix.exe -d D:\WSL\Fedora29
LXRunOffline.exe install help
LXRunOffline.exe config-uid -n Fedora -v <user_uid>
Set default user for Fedora as above in the command prompt. (You can get the user id by running the id command in Fedora Terminal)
On first startup, you will be prompted to create a username and password. Note that each WSL instance will have it's own useraname and password that are not synced with the Windows user.
You can start the distro from the Ubuntu icon on the Start Menu, or by running wsl or bash from a PowerShell or CMD console. You can go into fullscreen on WSL/CMD/PowerShell (native consoles or also in Windows Terminal sessions) with Alt-Enter. Registered distros are automatically added to Windows Terminal (how to add a new Windows Terminal console type?).
Remove scrollbar in fullscreen
You can also execute a number of Linux commands without having to first launch into the dedicated shell. This is handy for quick processes, for example running an update.
To do this you would use the template wsl <argument> <options> <commandline>
To run commands in your default Linux distro, you don't need to specify an argument at all.
e.g. To update Ubuntu if it's your default you would simply enter wsl sudo apt update
wsl --distribution debian sudo apt update
wsl -d debian sudo apt install youtube-dl -y
What about creating a user/pass on a new instance?
It is important to understand that there is no systemd in WSL, so distros cannot use standard shutdown/reboot or any other actions dependent upon systemd. Closing a WSL window will not shutdown the WSL engine. The WSL instance (and any services, websites, containers) continue to run in the background. Only terminating the WSL distro will stop those services. More on that here.
Shutdown and Reboot: There are times when it is important to shutdown or reboot the full WSL instance for configuration changes, hence the following commands have been added to WSL section of .custom (i.e. this section is only run if a distro is detected to be WSL) to replace the built-in (systemd) shutdown and reboot:
alias shutdown='cmd.exe /c "wsl.exe -t $WSL_DISTRO_NAME"'
alias reboot='cd /mnt/c/ && cmd.exe /c start "Rebooting WSL ..." cmd /c "timeout 5 && title "$WSL_DISTRO_NAME" wsl -d $WSL_DISTRO_NAME" && wsl.exe --terminate $WSL_DISTRO_NAME'
Reset a WSL distro back to an initial state
Settings > Apps > Apps & features > select the Linux Distro Name
In the Advanced Options link, select the "Reset" button to restore back to initial install state (note that everything will be deleted in that distro!) which is very useful for testing distros.
WSL images run in Hyper-V images via the LxssManager service. Therefore, to restart all WSL instances, just restart the service Get-Service LxssManager | Restart-Service.
Upgrade a WSL distro from WSL 1 to WSL 2 with wsl --set-version Ubuntu 2 (cannot be undone after upgrade)
Set the default distro with wsl --setdefault Ubuntu (it will now start when wsl or bash are invoked from DOS/PowerShell).
e.g. if you now run a command that uses wsl.exe it will use the default distro: Get-Process | wsl grep -i Win
From PowerShell, list current images with wsl -l -v ( --list --verbose ):
NAME STATE VERSION
* fedoraremix Running 1
* Debian Running 1
Ubuntu-20.04 Running 2
WSL gives seamless access between WSL and the Windows filesystem, including opening/editing files in WSL with Windows tools (like Notepad++ or VS Code), and opening/editing files in Windows with Linux tools.
VS Code: The WSL VS Code extension loads a VS Code server inside WSL that handles seamless integration (~/.vscode-server) which can be around 100 MB. If the
Alternatively, it is possible to just navigate to \\wsl$\Ubuntu\home\<user> even when the WSL window is closed (but NOT if the session is terminated).
Setup 'Quick access' link to the WSL distro's home folder: Open Windows Explorer, then navigate to \\wsl$\Ubuntu\home\<user>. Drag this folder into 'Quick access' for eacy access from Windows Explorer to see files in the distro home folder.
If choco install lxrunoffline is installed, from Windows Explorer, note the "LxRunOffline" right-click item to let you open a bash shell at that folder location with the chosen distro.
From inside a WSL shell, to open the current folder in Windows Explorer, use explorer.exe .
Running any Windows Executable from WSL Linux
Some aliases can make working with WSL in Windows easier (some templates have been added to .custom):
alias start=explorer.exe # "start ." will now open Explorer at current folder, same as "start ." in DOS/PowerShell
alias chrome="\"/mnt/c/Program Files/Google/Chrome/Application/chrome.exe\"" # Chrome. Can use with URL: chrome www.google.com
alias notepad++="\"/mnt/c/Program Files/Notepad++/notepad++.exe\"" # Notepad++. Can use with files: notepad++ ~/.bashrc
Opening a file with a Windows tool as above uses the \\wsl$ share, e.g. .bashrc in Linux will displays as \\wsl$\Ubuntu-20.04\home\boss\.bashrc in Windows Explorer. The Linux home folder ~ maps to \\wsl$\Ubuntu-20.04\home\boss (or \\wsl$\Ubuntu-20.04\root for root) and not to %userprofile%. \\wsl$ does not display in net share and you cannot run cd \\wsl$ or dir \\wsl$ but you can use the distro name, cd \\wsl$\Ubuntu-20.04 or dir \\wsl$\Ubuntu-20.04\home\boss etc. You can navigate to \\wsl$ in Windows Explorer however, and pin it or any subfolder to 'Quick access'.
VS Code Remote WSL Extension
WSL integration enables you to store your project files on the Linux file system using Linux command line tools, but also edit, debug, run projects directly in VS Code on Windows (or in a Chrome browser one Windows, or Notepad++ etc) generally without any performance issues associated with running across Linux and Windows file systems.
WSL Console Notes
To use Ctrl+Shift+C / Ctrl+Shift+V for Copy/Paste operations in the console, you need to enable the "Use Ctrl+Shift+C/V as Copy/Paste" option in the Console “Options” properties page (done this way to ensure not breaking any existing behaviors).
WSL Command References: wsl.exe / wslconfig.exe / bash.exe
Run bash.exe (or wsl.exe) from a cmd prompt to launch linux from the current working directory and using the default distro. Run bash ~ (or wsl ~) would launch in the user's home directory. Can start in any folder in this way, so wsl /mnt/c/Users/John would start in the Windows home folder of John. wsl -d Fedoraremix ~ would open in a specific distro version.
Keep in mind that:
• PowerShell is case-insensitive and uses objects by default.
• Linux is case-sensitive and passes strings by default.
Examples from a PowerShell prompt. As above, not that these are case-insensitive by default.
Get processes with "win" in the name (the following fails due to requiring to convert to a string first):
PS C:\> Get-Process | Select-String win
Fixing this by converting the object to a string first.
As Select-String will try to use objects from the pipe, Out-String -Stream is required to flatten that to a string.
PS C:\> Get-Process | Out-String -Stream | Select-String win
Using Get-Alias -Definition <cmdlet> to find the aliases for each Cmdlet:
PS C:\> ps | Out-String -Stream | sls win
Now do the same but using grep and sed from WSL to manipulate the output.
Note that when piping to a non-PowerShell command (not just WSL, but anything non-PowerShell), PowerShell automatically flattens that output to a string so Out-String -Stream is not required. Also note that ps here is Get-Process from PowerShell, and not ps from the WSL Linux distro.
PS C:\> ps | grep -i win | wsl sed 's/win/xxx/g'
We can change the default distro using -s, so could use multiple different WSL instances on a single line:
PS C:\> Write-Output "Hello``nLinux" | wsl -d Ubuntu grep -i linux | wsl -d Debian cowsay
Remember that the PowerShell is case-insensitive but that the Linux is case-sensitive so get different output without the -i.
PS C:\> ps | wsl grep win
Making this case-insensitive, so will be equivalent to ps | out-string -stream | sls win
PS C:\> ps | wsl grep -i win
A more complex example, to access the Windows filesystem (a more complex example). winhome in this example searches for and stores the case-sensitive User folder name from the Windows filesystem (due to case-sensitivity in Linux, this folder could have different case than the Linux user name. e.g. "C:\Users\John" in Windows might be "/home/john" in WSL):
$ winhome=$(find /mnt/c/Users -maxdepth 1 -type d -regextype posix-extended -iregex /mnt/c/users/$USER)
Here are some notes on differences between /mnt/ host (Windows) drive mountpoints and the Linux filesystem.
Non-Microsoft open source tool to manage WSL (there was a now deprecated tool called lxrun.exe that was previously part of WSL that this tool is named after). Home page), Git Project.
"A full-featured utility for managing Windows Subsystem for Linux (WSL)"
Chocolatey: choco install lxrunoffline
Scoop: scoop bucket add extras, scoop install lxrunoffline
Shell extension: The right-click menu feature requires the shell extension DLL to be properly registered. This is automatically done if you used Chocolatey to install this project. However, if you downloaded the binaries directly, you need to run regsvr32 LxRunOfflineShellExt.dll manually to register the DLL file.
PS C:\> LxRunOffline.exe
Supported actions are:
l, list List all installed distributions.
gd, get-default Get the default distribution, which is used by bash.exe.
sd, set-default Set the default distribution, which is used by bash.exe.
i, install Install a new distribution.
ui, uninstall Uninstall a distribution.
rg, register Register an existing installation directory.
ur, unregister Unregister a distribution but not delete the installation directory.
m, move Move a distribution to a new directory.
d, duplicate Duplicate an existing distribution in a new directory.
e, export Export a distribution's filesystem to a .tar.gz file, which can be imported by the "install" command.
r, run Run a command in a distribution.
di, get-dir Get the installation directory of a distribution.
gv, get-version Get the filesystem version of a distribution.
ge, get-env Get the default environment variables of a distribution.
se, set-env Set the default environment variables of a distribution.
ae, add-env Add to the default environment variables of a distribution.
re, remove-env Remove from the default environment variables of a distribution.
gu, get-uid Get the UID of the default user of a distribution.
su, set-uid Set the UID of the default user of a distribution.
gk, get-kernelcmd Get the default kernel command line of a distribution.
sk, set-kernelcmd Set the default kernel command line of a distribution.
gf, get-flags Get some flags of a distribution. See https://docs.microsoft.com/en-us/previous-versions/windows/desktop/api/wslapi/ne-wslapi-wsl_distribution_flags for details.
sf, set-flags Set some flags of a distribution. See https://docs.microsoft.com/en-us/previous-versions/windows/desktop/api/wslapi/ne-wslapi-wsl_distribution_flags for details.
s, shortcut Create a shortcut to launch a distribution.
ec, export-config Export configuration of a distribution to an XML file.
ic, import-config Import configuration of a distribution from an XML file.
sm, summary Get general information of a distribution.
version Get version information about this LxRunOffline.exe.
https://stackoverflow.com/questions/38779801/move-wsl-bash-on-windows-root-filesystem-to-another-hard-drive In any Windows 10 version, you can move the distribution to another drive using lxRunOffline.
Default locations for WSL Distros
C:\Program Files\WindowsApps\CanonicalGroupLimited.Ubuntu20.04onWindows_2004.2020.812.0_x64__79rhkp1fndgsc
-
Set permissions to the target folder. First, I think you must set some permissions to the folder where the distribution will be moved. Use
icaclsto set the proper permissions.
C:\> whoami
test\john
C:\> icacls D:\wsl /grant "john:(OI)(CI)(F)"
NOTE: In addition to the above permissions, I have activated the long path names in Windows. -
Move the distribution with
lxrunoffline move.
C:\wsl> lxrunoffline move -n Ubuntu-18.04 -d d:\wsl\installed\Ubuntu-18.04
You may check the installation folder using
C:\wsl> lxrunoffline get-dir -n Ubuntu-18.04
d:\wsl\installed\Ubuntu-18.04 -
Run the distribution. after moving the distribution, you can run the distribution using
wslor the samelxrunoffline
C:\wsl> lxrunoffline run -n Ubuntu-18.04 -w
user@test:~$ exit
logout
C:\wsl> wsl
user@test:/mnt/c/wsl$ exit
logout
Git direct passwords were deprecated on 21st August 2021, so need to move over to tokens.
WSL-Git
Use WSL Git instead of Git for Windows (StackOverflow)
Git
Set your email as follows command (setting name and email you use on your Git account):
git config --global user.name "username"
git config --global user.email "[email protected]"
Git Credential Manager setup
Git Credential Manager enables you to authenticate a remote Git server, even if you have a complex authentication pattern like two-factor authentication, Azure Active Directory, or using SSH remote URLs that require an SSH key password for every git push. Git Credential Manager integrates into the authentication flow for services like GitHub and, once you're authenticated to your hosting provider, requests a new authentication token. It then stores the token securely in the Windows Credential Manager. After the first time, you can use git to talk to your hosting provider without needing to re-authenticate. It will just access the token in the Windows Credential Manager.
I have had some failures getting /usr/bin/git to work. A quick way that always works, if you don't need more than the git command on its own, is to just alias git.exe from Git for Windows from WSL and just use that for managing repos.
alias git="'/mnt/c/Program Files/Git/bin/git.exe'" (note that "' / '" are required for this alias). Also make sure that the following section is in ~/.gitconfig:
[credential]
helper = manager
With the above setup, you do not need git installed in WSL at all.
If you want git in WSL (/usr/bin/git) to use the Git for Windows Credential Manager however, this guide explains the steps:
• Install Git for Windows.
• In Windows-land, clone any single private repo to generate the token. This will open a GitHub dialog allowing you to log in, and this will also handle 2FA (two-factor authentication) correctly.
• In your WSL distro, run sudo vi /usr/bin/git-credential-manager, then add the following lines:
#!/bin/sh
exec "/mnt/c/Program\ Files/Git/mingw64/libexec/git-core/git-credential-manager.exe" $@
Alternatively to the above, the following command should achieve the same thing:
git config --global credential.helper "/mnt/c/Program\ Files/Git/mingw64/libexec/git-core/git-credential-manager.exe"
• Run sudo chmod +x /usr/bin/git-credential-manager
Now, open up ~/.gitconfig and add the following section:
[credential]
helper = manager
• Make sure you clone all your repos with https:// URLs. You do not need www., just github.com, and you do not need to put .git at the end of a project URL, just use the URL for the project page) or re-configure them via git remote set-url.
This is done as WSL has no GUI, so can't run the credential manager inside the WSL Linux VM. Now any git operation you perform within your WSL distribution should use the credential manager token from Windows seamlessly. If you already have credentials cached for a host, it will access them from the credential manager. If not, you'll receive a dialog response requesting your credentials, even if you're in a Linux console. If you are using a GPG key for code signing security, you may need to associate your GPG key with your GitHub email.
Git Ignore file
It is usually important to add .gitignore files to most projects (must be in the root folder of a project) to exclude certain files. GitHub offers a collection of useful .gitignore templates with recommended .gitignore file setups organized according to your use-case. For example, here are GitHub's default .gitignore templates. If you choose to create a new repo using the GitHub website, there are check boxes available to initialize your repo with a README file, .gitignore file set up for your specific project type, and options to add a license if you need one.
Add one line per exclude mask (can use wildcards). If a file is already in the project, it must be removed:
git rm --cached <filename>
To create a global ignore file (e.g. could put *.tmp to exclude all files like that in all projects):
git config --global core.excludesfile ~/.gitignore_global
To exclude without using .gitignore, put exclude masks into .git/info/exclude.
Info on both Git and Docker with WSL
Note that the above contains the following comment "Apparently WSL has kinda crappy IO performance". This comment relates to WSL 1 only. WSL 2 is about 15x to 20x faster for IO operations due to the native Linux kernel build into the OS for WSL 2 etc.
(Should I use docker or Docker-Desktop or both??)
Linux disk images install by default to the C: drive.
In Windows Powershell, run wsl --list to view Linux distros.
Export current distro: wsl --export Ubuntu D:\Backups\Ubuntu.tar
Unregister this distro: wsl --unregister Ubuntu
Run wsl --list to verify the distribution has been removed.
Import backup to new WSL distro at new location: wsl --import Ubuntu D:\WSL\ D:\Backups\Ubuntu.tar
Run wsl --list to verify that it has been created and registered.
Launch the distro as normal from the Start menu.
Unfortunately, Ubuntu will now use root as the default user. To go back to your own account ubuntu config --default-user <yourname> where <yourname> is the username you defined during installation.
Multiple instances of same Linux distro in WSL
To use Ctrl+Shift+C/V for Copy/Paste operations in the console, need to enable the "Use Ctrl+Shift+C/V as Copy/Paste" option in the Console “Options” properties page (done this way to ensure not breaking any existing behaviors).
Note that doing this will break all connections to open sessions, e.g. open VS Code editor sessions to scripts inside WSL.
When you restart the instance, VS Code sessions will not reconnect, but just close VS Code (it will cache the scripts). When you restart
VS Code (e.g. code .) against the same files, VS Code will reopen with a connection to the scripts and you can continue.
You can terminate a single WSL instance from PowerShell as follows (will terminate the Hyper-V VN immediately):
wsl -t <distro-name> # or --terminate
You can also terminate a WSL session from its own console as WSL imports the whole Windows PATH so that you can run all Windows binaries, but you have to fully specify the name of a binary (so must use wsl.exe and not just wsl), so to terminate a WSL session, use:
wsl.exe -t <distro-name>
Note that sudo reboot does not work (as it is part of systemd which does not exist on any WSL distro). It will generate the following error:
System has not been booted with systemd as init system (PID 1). Can't operate.
Failed to connect to bus: Host is down
Failed to talk to init daemon.
All systemd actions will fail, e.g. systemctl start etc. More on this here.
Also look at echo $PATH to see that all Windows paths are imported (this happens at session startup).
Kill all WSL instances: Get-Service LxssManager | Stop-Service (then same again but with | Start-Service to start again)
Restart all instances: Get-Service LxssManager | Restart-Service
WSL2-Hacks https://stackoverflow.com/questions/55579342/why-systemd-is-disabled-in-wsl
To make all new distros install as v2:
wsl --set-default-version 2
To upgrade an existing distro to v2:
wsl --set-version <distro-name> 2
Conversion in progress, this may take a few minutes...
For information on key differences with WSL 2 please visit https://aka.ms/wsl2
You cannot use sudo reboot or any similar systemd command:
Systemd command Sysvinit command
systemctl start service_name service service_name start
systemctl stop service_name service service_name stop
systemctl restart service_name service service_name restart
systemctl status service_name service service_name status
systemctl enable service_name chkconfig service_name on
systemctl disable service_name chkconfig service_name off
Why you probably don't need systemd on WSL
Most popular Linux distributions use systemd as the init system for startup, shutdown, service monitoring, etc, but WSL has it's own initialization system, and so no WSL distros use systemd, and do not generally employ a traditional init system.
Two skills in particular will yield more satisfaction with WSL:
• A good understanding of how to launch services directly (unmanaged by an init system).
• Familiarity with running containers. Without docker. (If you use WSL version 2).
How to install systemd snap packages on WSL 2
sudo apt update
sudo apt install -yqq git
git clone https://github.com/DamionGans/ubuntu-wsl2-systemd-script.git $HOME/ubuntu-wsl2-systemd-script
$HOME/ubuntu-wsl2-systemd-script/ubuntu-wsl2-systemd-script.sh
cd $HOME/ubuntu-wsl2-systemd-script
git pull
./ubuntu-wsl2-systemd-script.sh --force
Connect to WSL via SSH
SSH into a WSL2 host remotely and reliably
sudo apt install openssh-server # Install SSH server
/etc/ssh/sshd_config # Change Port 22 to Port 2222 as Windows uses port 22
sudo visudo # We setup service ssh to not require a password
# Allow members of group sudo to execute any command
%sudo ALL=(ALL:ALL) ALL
%sudo ALL=NOPASSWD: /usr/sbin/service ssh *
sudo service ssh --full-restart # Restart ssh service sudo /etc/init.d/ssh start
You might see: sshd: no hostkeys available -- exiting
If so, you need to run: sudo ssh-keygen -A to generate in /etc/ssh/
Now restart the server: sudo /etc/init.d/ssh start
You might see the following error on connecting: "No supported authentication methods available (server sent: publickey)"
To fix this, sudo vi /etc/ssh/sshd_config. Change as follows to allow username/password authentication:
PasswordAuthentication = yes
ChallengeResponseAuthentication = yes
Restart ssh sudo /etc/init.d/ssh restart (or sudo service sshd restart).
Note: If you set PasswordAuthentication to yes and ChallengeResponseAuthentication to no you are able to connect automatically with a key, and those that don't have a key will connwct with a password - very useful
PuttyGen will create a public key file that looks like:
---- BEGIN SSH2 PUBLIC KEY ----
Comment: "rsa-key-20121022"
AAAAB3NzaC1yc2EAAAABJQAAAIEAhGF6GIuMY8FJ1+CNApnSY1N2YSlkYz72Yvwu
a6N1nFpBklz1+dsIMg4rcTLcF34M/tW5Yz+NUDAw2AEbxQ32FPgw7sAOIXktkYOH
tr7mmimiTjkoSCrJh1kqalPSpi8rglT/Bp67Ql2SZwvUFfMzHISryR0EZC4rXP/u
vObrJe8=
---- END SSH2 PUBLIC KEY ----
However, this will not work, so what you need to do is to open the key in PuttyGen, and then copy it from there (this results in the key being in the right format and in 1 line):
ssh-rsa AAAAB3NzaC1yc2EAAAABJQAAAIEAhGF6GIuMY8FJ1+CNApnSY1N2YSlkYz72Yvwua6N1nFpBklz1+dsIMg4rcTLcF34M/tW5Yz+NUDAw2AEbxQ32FPgw7sAOIXktkYOHtr7mmimiTjkoSCrJh1kqalPSpi8rglT/Bp67Ql2SZwvUFfMzHISryR0EZC4rXP/uvObrJe8= rsa-key-20121022
Paste this into authorized_keys then it should work.
I also try use it as a remote gdb server for visual studio by VisualGDB,it not works well. VisualGDB will support it in the next version as the offical website shows.The link is https://sysprogs.com/w/forums/topic/visualgdb-with-windows-10-anniversary-update-linux-support/#post-9274
Enable ability to connect to WSL locally on 127.0.0.1 (port 2222) or remotely from any system in the network
We need to make sure that OpenSSH Server is installed and a few settings are changed in the /etc/ssh/sshd_config file.
sudo apt install openssh-server # But it is probably already installed
Now run sudo vi /etc/ssh/sshd_config :
Port 2222 # Default is '22', must change for WSL
PermitRootLogin prohibit-password # Change?
PasswordAuthentication yes # Default is 'no'
AllowAgentForwarding yes
X11Forwarding yes
We must make sure that key-pairs are generated (for root, these generate in /etc/ssh and for the current user, they generate in /home/<username>/.ssh by default):
sudo ssh-keygen -A # Generate root keys
ssh-keygen -t rsa -b 4096 # Generate a 4096 bits SSH key pair for current user
Now, we need to restart the ssh service, but need to remember that WSL does not use systemd, so sudo systemctl restart sshd.service will not work (neither will sudo reboot etc for the same reason):
sudo service --status-all # Show all init.d service names (look for "ssh")
sudo /etc/init.d/ssh stop # Stop/Start the ssh service after every change
sudo service ssh start # This will fail if keys are not generated
sudo service --status-all # Can now see that the serice is running
Can now connect LOCALLY via putty or other ssh client at 127.0.0.1 on port 2222 (remember that Windows now has a built in ssh.exe and ssh-keygen.exe etc).
To connect from a REMOTE system to this instance of WSL, follow this guide:
edit /etc/ssh/sshd_config with the following three changes
Port 2222
ListenAddress 0.0.0.0
PasswordAuthentication yes
We also need to edit /etc/sudoers.d/ in order to remove the requirement of a password for starting the ssh service, this will come handy later on in the automation section of the article, so add the following line:
%sudo ALL=NOPASSWD: /usr/sbin/service ssh *
After all this we can start the service:
service ssh start
Starting WSL with keyboard shortcut ...
How to install Fedora Remix 33 for WSL 2 ...
Guide to using Linux GUI apps + ZSH + Docker on WSL 2
Nicky Meuleman’s guide on Using Graphical User Interfaces like Cypress’ in WSL2
Nicky Meuleman’s guide on Linux on Windows WSL 2 + ZSH + Docker
Notes on Zsh and Oh My Zsh
Nicky Meuleman’s guide on setting up ZSH
There is talk of the WSL team Adding Linux GUI app support to WSL but this is slated for an update around December 2020, but this will only be in the Insider's Build and probably not in the main release of Windows 10 until the Spring/Summer 2021 release.
Support for Wayland GUI Apps in WSL 2
With the latest updates to WSL2, Linux GUI apps integrate with Windows 10 using Wayland display server protocol running inside of WSL. Wayland communicates with a Remote Desktop Protocol (RDP) client on the Windows host to run the GUI app.
Microsoft is also bringing GPU hardware acceleration for Linux applications running in WSL2. It has pushed the first draft of the brand new GPU driver ‘Dxgkrnl’ for the Linux kernel.
Run Windows commands from WSL and vice-versa
In the Windows 10 Creators Update (build 1703, April 2017), this is natively supported. So you can now run Windows binaries from Linux...
alias putty="/mnt/c/ProgramData/chocolatey/bin/PUTTY.EXE" # Run Windows commands form Linux
bash -c "fortune | cowsay" # Run Linux commands from Windows PowerShell
Quick summary
Reboot: do not use sudo reboot, use wsl -t <distro> (or from inside the session itself, use wsl.exe -t <distro>)
To shutdown all distros immediately (i.e. to close the WSL VM manager): wsl -shutdown (or from inside the session itself, use wsl.exe -shutdown)
This is very important for seamless integration between WSL sessions and Windows, so that you can open any script on the WSL session using VS Code. Install this extension from within VS Code to be able to run code ~/.bashrc directly inside a WSL session and have that open in VS Code.
... add more notes here ...
WSL Links
WSL: The Ultimate Guide
Linux Graphical Apps coming to WSL
Target WSL from Visual Studio
Fun with WSL (older page, some info is out of date)
Custom Functions for Profiles
Large list
6 useful one-liners
10 useful one-liners for system
Large list of lesser known commands
Linux Complex Bash One-Liner Examples
Random
Random
Interesting Vim & Terminal colour notes
Uses of the ! command
Fun Stuff 1
Fun Commands I
Fun Commands II
Fun with Character Counts
https://stackoverflow.com/questions/36585496/error-when-using-git-credential-helper-with-gnome-keyring-as-sudo/40312117#40312117
netstat -plan | grep :80 | awk '{print $5}' | cut -d: -f1 | sort | uniq -c | sort -nk1
Above command will give a sorted list of IP’s with number of connections to port 80.
https://aka.ms/wsl
https://docs.microsoft.com/en-us/windows/wsl/wsl-config#list-distributions
WSL 2 architecture
Traditional VMs are slow to boot up, isolated, and consume a lot of resources. WSL 2 does not have these attributes, and requires no VM configuration or management While WSL 2 does use a VM, it is managed and run behind the scenes, leaving you with the same user experience as WSL 1.
Full Linux kernel
The Linux kernel in WSL 2 is built by Microsoft from the latest stable branch, based on the source available at kernel.org. This kernel has been specially tuned for WSL 2, optimizing for size and performance, and will be serviced by Windows updates for latest security fixes and kernel improvements.
Increased file IO performance
File intensive operations like git clone, npm install, apt update, apt upgrade, and more are all noticeably faster with WSL 2. Initial versions of WSL 2 run up to 20x faster compared to WSL 1 when unpacking a zipped tarball, and around 2-5x faster when using git clone, npm install and cmake on various projects.
Full system call compatibility
Linux binaries use system calls to perform functions such as accessing files, requesting memory, creating processes, and more. Whereas WSL 1 used a translation layer that was built by the WSL team, WSL 2 includes its own Linux kernel with full system call compatibility. Benefits include:
• A whole new set of apps that you can run inside of WSL, such as Docker and more.
• Updates to the Linux kernel are immediately ready for use (You don't have to wait for the WSL team to implement updates and add the changes).
Accessing network applications
Accessing Linux networking apps from Windows (localhost)
If you are building a networking app (for example an app running on a NodeJS or SQL server) in your Linux distribution, you can access it from a Windows app (like your Edge or Chrome internet browser) using localhost (just like you normally would).
To find the IP address of the virtual machine powering your Linux distribution:
From your WSL distribution (ie Ubuntu), run the command: ip addr Find and copy the address under the inet value of the eth0 interface. If you have the grep tool installed, find this more easily by filtering the output with the command: ip addr | grep eth0 Connect to your Linux server using this IP address. The picture below shows an example of this by connecting to a Node.js server using the Edge browser.
Connect to NodeJS server with Edge
Accessing Windows networking apps from Linux (host IP) If you want to access a networking app running on Windows (for example an app running on a NodeJS or SQL server) from your Linux distribution (ie Ubuntu), then you need to use the IP address of your host machine. While this is not a common scenario, you can follow these steps to make it work.
Obtain the IP address of your host machine by running this command from your Linux distribution: cat /etc/resolv.conf Copy the IP address following the term: nameserver. Connect to any Windows server using the copied IP address. The picture below shows an example of this by connecting to a Node.js server running in Windows via curl.
Connect to NodeJS server in Windows via Curl
Additional networking considerations Connecting via remote IP addresses When using remote IP addresses to connect to your applications, they will be treated as connections from the Local Area Network (LAN). This means that you will need to make sure your application can accept LAN connections.
For example, you may need to bind your application to 0.0.0.0 instead of 127.0.0.1. In the example of a Python app using Flask, this can be done with the command: app.run(host='0.0.0.0'). Please keep security in mind when making these changes as this will allow connections from your LAN.
Accessing a WSL 2 distribution from your local area network (LAN) When using a WSL 1 distribution, if your computer was set up to be accessed by your LAN, then applications run in WSL could be accessed on your LAN as well.
This isn't the default case in WSL 2. WSL 2 has a virtualized ethernet adapter with its own unique IP address. Currently, to enable this workflow you will need to go through the same steps as you would for a regular virtual machine. (We are looking into ways to improve this experience.)
Here's an example PowerShell command to add a port proxy that listens on port 4000 on the host and connects it to port 4000 to the WSL 2 VM with IP address 192.168.101.100.
PowerShell
Copy netsh interface portproxy add v4tov4 listenport=4000 listenaddress=0.0.0.0 connectport=4000 connectaddress=192.168.101.100 IPv6 access WSL 2 distributions currently cannot reach IPv6-only addresses. We are working on adding this feature.
Expanding the size of your WSL 2 Virtual Hard Disk WSL 2 uses a Virtual Hard Disk (VHD) to store your Linux files. In WSL 2, a VHD is represented on your Windows hard drive as a .vhdx file.
The WSL 2 VHD uses the ext4 file system. This VHD automatically resizes to meet your storage needs and has an initial maximum size of 256GB. If the storage space required by your Linux files exceeds this size you may need to expand it. If your distribution grows in size to be greater than 256GB, you will see errors stating that you've run out of disk space. You can fix this error by expanding the VHD size.
To expand your maximum VHD size beyond 256GB:
Terminate all WSL instances using the command: wsl --shutdown
Find your distribution installation package name ('PackageFamilyName')
Using PowerShell (where 'distro' is your distribution name) enter the command: Get-AppxPackage -Name "" | Select PackageFamilyName Locate the VHD file fullpath used by your WSL 2 installation, this will be your pathToVHD:
%LOCALAPPDATA%\Packages<PackageFamilyName>\LocalState<disk>.vhdx Resize your WSL 2 VHD by completing the following commands:
Open Windows Command Prompt with admin privileges and enter:
PowerShell
Copy diskpart DISKPART> Select vdisk file="" DISKPART> detail vdisk Examine the output of the detail command. The output will include a value for Virtual size. This is the current maximum. Convert this value to megabytes. The new value after resizing must be greater than this value. For example, if the detail output shows Virtual size: 256 GB, then you must specify a value greater than 256000. Once you have your new size in megabytes, enter the following command in diskpart:
PowerShell
Copy DISKPART> expand vdisk maximum= Exit diskpart
PowerShell
Copy DISKPART> exit Launch your WSL distribution (Ubuntu, for example).
Make WSL aware that it can expand its file system's size by running these commands from your Linux distribution command line.
Note
You may see this message in response to the first mount command: /dev: none already mounted on /dev. This message can safely be ignored.
PowerShell
Copy sudo mount -t devtmpfs none /dev mount | grep ext4 Copy the name of this entry, which will look like: /dev/sdX (with the X representing any other character). In the following example the value of X is b:
PowerShell
Copy sudo resize2fs /dev/sdb M Note
You may need to install resize2fs. If so, you can use this command to install it: sudo apt install resize2fs.
The output will look similar to the following:
Bash
Copy resize2fs 1.44.1 (24-Mar-2018) Filesystem at /dev/sdb is mounted on /; on-line resizing required old_desc_blocks = 32, new_desc_blocks = 38 The filesystem on /dev/sdb is now 78643200 (4k) blocks long. Note
In general do not modify, move, or access the WSL related files located inside of your AppData folder using Windows tools or editors. Doing so could cause your Linux distribution to become corrupted.
Enable root user in various distros
• For Ubuntu:
ubuntu config --default-user <user_name>
• For openSuse:
opensuse-42 config --default-user <user_name>
• For SUSE Linux Enterprise Server:
sles-12 config --default-user <user_name>
Use root for a spcific task:
sudo -i
To enable the root account:
sudo -i passwd root
su - (or su - root)
To disable the root account:
sudo passwd -dl root
To enable the root account login through the graphical user interface, we need to edit the next two files:
/etc/gdm3/custom.conf AllowRoot=true
/etc/pam.d/gdm-password
Comment out the line auth reuireqd pam_succeed_if.so user != root quiet_access
https://code.visualstudio.com/docs/remote/wsl
https://docs.microsoft.com/en-us/windows/wsl/tutorials/wsl-vscode
https://ajeet.dev/developing-in-wsl-using-visual-studio-code/
Install a distro from App Store or Chocolatey or manually with iwr/curl
• App Store: Distros https://aka.ms/wslstore
• Chocolatey wsl-ubuntu-2004, wsl-fedoraremix, wsl-alpine
choco install wsl-ubuntu-2004
This chocolatey package can set the default user as root (default is false):
choco install wsl-ubuntu-2004 --params "/InstallRoot:true"
• Using iwr/curl (Invoke-WebRequest):
iwr -Uri https://aka.ms/wslubuntu2004 -OutFile Ubuntu.appx -UseBasicParsing
curl.exe -L -o Ubuntu.appx https://aka.ms/wslubuntu2004 # curl.exe
Finally, install and register with: Add-AppxPackage .\Ubuntu.appx # To install the AppX package
After doing this, wsl -l -v will show the registered distro.
To inject into the configurations files (.bashrc, .vimrc, .intputrc, sudoers), custom_loader.sh uses a few grep, sed, and tee techniques to selectively remove only specific lines and then replace them. For .bashrc, the goal is to make sure that the .custom lines are always positioned at the end of file so that this only loads after .bashrc so that this project makes sure to alter core files as little as possible.
Using grep to prune lines
First define the exact line to be pruned (I could use a smaller subset like [ ! -f ~/.custom ] as that would be enough, since it relates to .custom so can only be relating to this project, but in this case, we prune the entire line).
GETCUSTOM='[ ! -f ~/.custom ] && [[ $- == *"i"* ]] && curl -s https://raw.githubusercontent.com/roysubs/custom_bash/master/.custom > ~/.custom'
Then search only for this specific line (-q silent show no output, -x full line match, -F fixed string / no regexp), and then appends directly onto the file (have to use tee here as you cannot > directly onto the file as that operation will happen while the grep is happening; by using tee the entire grep will finish then append to the file).
grep -qxF "$GETCUSTOM" ~/.bashrc || echo $HEADERCUSTOM | tee --append ~/.bashrc
Note: if doing this with a system file (like /etc/bashrc), you have to also sudo the tee command:
grep -qxF "$GETCUSTOM" /etc/bashrc || echo $HEADERCUSTOM | sudo tee --append ~/.bashrc
This is because even though you might be able to read /etc/bashrc, writing to it must be elevated, and note that putting sudo before the grep command will not do that (elevates only the grep process and not the tee process).
Removing empty lines but only from the end of a file
This is to make sure no empty lines at end of file, except for one required before our lines.
sed -i -e :a -e '/^\n*$/{$d;N;};/\n$/ba' ~/.bashrc # Removes also any empty lines from the end of a file. https://unix.stackexchange.com/questions/81685/how-to-remove-multiple-newlines-at-eof/81687#81687
echo "" | tee --append ~/.bashrc # Add an empty line back in as a separator before our the lines to call .custom
.inputrc changes.
set completion-ignore-case On really useful to prevent case-sensitivity taking over when trying to tab complete into folders
.vimrc
sudoers
Notes on changes to sudoers here ...
The Ultimate Guide to Windows Subsystem for Linux (Windows WSL) 2021-02-02 https://adamtheautomator.com/windows-subsystem-for-linux/ (END)