wsl2-hacks

Useful snippets / tools for using WSL2 as a development environment

---

Auto-start/services (systemd and snap support)

I've done a few methods that have had various levels of success. My goal was to make it feel seamless for my workflow and have commands work as expected. What's below is the current version of the setup I use. It allows me to use the MS Terminal as well as VSCode's Remote WSL plugin.

With this setup your shells will be able to run systemctl commands, have auto-starting services, as well as be able to run snaps.

1. Install deps

   $ sudo apt update
   $ sudo apt install dbus policykit-1 daemonize

2. Create a fake-bash

   This fake shell will intercept calls to wsl.exe bash ... and forward them to a real bash running in the right environment for systemd. If this sounds like a hack-- well, it is. However, I've tested various workflows and use this daily. That being said, your mileage may vary.

   $ sudo touch /usr/bin/bash
   $ sudo chmod +x /usr/bin/bash
   $ sudo editor /usr/bin/bash
   

   Add the following, be sure to replace <YOURUSER> with your WSL2 Linux username

   #!/bin/bash
   # your WSL2 username
   UNAME="<YOURUSER>"
   
   UUID=$(id -u "${UNAME}")
   UGID=$(id -g "${UNAME}")
   UHOME=$(getent passwd "${UNAME}" | cut -d: -f6)
   USHELL=$(getent passwd "${UNAME}" | cut -d: -f7)
   
   if [[ -p /dev/stdin || "${BASH_ARGC}" > 0 && "${BASH_ARGV[1]}" != "-c" ]]; then
       USHELL=/bin/bash
   fi
   
   if [[ "${PWD}" = "/root" ]]; then
       cd "${UHOME}"
   fi
   
   # get pid of systemd
   SYSTEMD_PID=$(pgrep -xo systemd)
   
   # if we're already in the systemd environment
   if [[ "${SYSTEMD_PID}" -eq "1" ]]; then
       exec "${USHELL}" "$@"
   fi
   
   # start systemd if not started
   /usr/sbin/daemonize -l "${HOME}/.systemd.lock" /usr/bin/unshare -fp --mount-proc /lib/systemd/systemd --system-unit=basic.target 2>/dev/null
   # wait for systemd to start
   while [[ "${SYSTEMD_PID}" = "" ]]; do
       sleep 0.05
       SYSTEMD_PID=$(pgrep -xo systemd)
   done
   
   # enter systemd namespace
   exec /usr/bin/nsenter -t "${SYSTEMD_PID}" -m -p --wd="${PWD}" /sbin/runuser -s "${USHELL}" "${UNAME}" -- "${@}"

3. Set the fake-bash as our root user's shell

   We need root level permission to get systemd setup and enter the environment. The way I went about solving this is to have WSL2 default to the root user and when wsl.exe is executed the fake-bash will do the right thing.

   The next step in getting this working is to change the default shell for our root user.

   Edit the /etc/passwd file:

   $ sudo editor /etc/passwd

   Find the line starting with root:, it should be the first line. Change it to:

   root:x:0:0:root:/root:/usr/bin/bash

   Note the /usr/bin/bash here, slight difference

   Save and close this file.

4. Exit out of / close the WSL2 shell

   The next step is to shutdown WSL2 and to change the default user to root.

   In a PowerShell terminal run:

   > wsl --shutdown
   > ubuntu1804.exe config --default-user root
   

5. Re-open WSL2

   Everything should be in place. Fire up WSL via the MS Terminal or just wsl.exe. You should be logged in as your normal user and systemd should be running

   You can test by running the following in WSL2:

   $ systemctl is-active dbus
   active

6. Create /etc/rc.local (optional)

   If you want to run certain commands when the WSL2 VM starts up, this is a useful file that's automatically ran by systemd.

   $ sudo touch /etc/rc.local
   $ sudo chmod +x /etc/rc.local
   $ sudo editor /etc/rc.local

   Add the following:

   #!/bin/sh -e
   
   # your commands here...
   
   exit 0

/etc/rc.local is only run on "boot", so only when you first access WSL2 (or it's shutdown due to inactivity/no-processes). To test you can shutdown WSL via PowerShell/CMD wsl --shutdown then start it back up with wsl.

---

Access localhost ports from Windows

NOTE: No longer needed as of build 18945

Many development servers default to binding to 127.0.0.1 or localhost. It can be cumbersome and frustrating to get it to bind to 0.0.0.0 to make it accessible via Windows using the IP of the WSL2 VM.

Take a look at https://github.com/shayne/go-wsl2-host to have wsl.local automatically resolve to the WSL2 VM

To make these dev servers / ports accessible you can run the following commands, or add them to the /etc/rc.local if you have systemd running:

# /etc/rc.local runs as root by default
# if you run these yourself add 'sudo' to the beginning of each command

$ sysctl -w net.ipv4.conf.all.route_localnet=1
$ iptables -t nat -I PREROUTING -p tcp -j DNAT --to-destination 127.0.0.1 

---

Increase max_user_watches

If devtools are watching for file changes, the default is too low.

# /etc/rc.local runs as root by default
# if you run these yourself add 'sudo' to the beginning of each command

sysctl -w fs.inotify.max_user_watches=524288

---

Open MS Terminal to home directory by default

Open your MS Terminal configuration Ctrl+,

Find the "commandLine":... config for the WSL profile.

Change to something like:

"commandline": "wsl.exe ~ -d Ubuntu-18.04",

Copy current IP of WSL2 into Windows clipboard (optionally with port 3000 here):

hostname -I | awk '{print $1}' | awk '{printf "%s:3000", $0}' | clip.exe

Alternatively, put it in a file, for example copy_ip.sh, make it executable with chmod +x copy_ip.sh and you can get the IP any time with ./copy_ip.sh:

#!/bin/bash
hostname -I | awk '{print $1}' | awk '{printf "%s:3000", $0}' | clip.exe