02. Tutorial for Rust development environment inside docker container. Rust - Hack Without Fear ! (docker_rust_development) (2022-03)
version: 2.0 date: 2022-04-13 author: repository: GitHub

Lines in md Lines in bash scripts

License Hits

Hashtags: #rustlang #buildtool #developmenttool #tutorial #docker #ssh
My projects on Github are more like a tutorial than a finished product: bestia-dev tutorials.


Trust in Rust open-source 2022

First I have to talk about TRUST. This is the whole point of this entire project.

We live in dangerous times for "supply chain attacks" in open-source and it is getting worse. This is a problem we need to address!
In this project, you don't need to TRUST ME! You can run all the bash commands inside bash scripts line-by-line. My emphasis is to thoroughly comment and describe what is my intention for every single command. You can follow and observe exactly what is going on. This is the beauty of open-source. But this is realistic only for very simple projects.
To be meticulously precise, you still have to trust the Windows code, Linux, GNU tools, drivers, Podman, Buildah, VSCode, extensions, the microprocessor, memory, disc and many other projects. There is no system without an inherent minimal level of TRUST.

Try it

Video tutorial on youtube:

Super short instructions without explanation just in 2 easy steps. For tl;dr; continue reading below. I tested the script on a completely fresh installation of Debian on WSL2.
Prerequisites: Win10, WSL2, VSCode and curl:

sudo apt update
sudo apt -y full-upgrade
sudo apt install -y curl

1. Prepare your existing SSH keys for github and publish-to-web, so you can use them from inside the container. Here we use the Win10 folder ~\.ssh because we want this data to be persistent and can survive the destruction of the container or even WSL2.

Download the templates for bash scripts and Then modify the scripts: replace the placeholders with your own data and file_names.
In WSL2 terminal:

mkdir $USERPROFILE/.ssh

curl -L -s --output $USERPROFILE/.ssh/

curl -L -s --output $USERPROFILE/.ssh/

# use sed to replace these words with your own data and file names:
# sed uses a strange syntax for replace 's/find_this/replace_with_this/g'
# 'info@your.mail', 'your_name', 'githubssh1' and 'webserverssh1' 

sed -i.bak 's/info@your.mail/info@my.mail/g' $USERPROFILE/.ssh/
sed -i.bak 's/your_name/my_name/g' $USERPROFILE/.ssh/
sed -i.bak 's/githubssh1/my_ssh1/g' $USERPROFILE/.ssh/
sed -i.bak 's/webserverssh1/my_ssh2/g' $USERPROFILE/.ssh/

sed -i.bak 's/githubssh1/my_ssh1/g' $USERPROFILE/.ssh/
sed -i.bak 's/webserverssh1/my_ssh2/g' $USERPROFILE/.ssh/

cat $USERPROFILE/.ssh/
cat $USERPROFILE/.ssh/

2. Download the podman_install_and_setup bash script and run it.
This can take 10 minutes to setup because it downloads 2 GB from docker hub.
In WSL2 terminal:

mkdir -p ~/rustprojects/docker_rust_development
cd ~/rustprojects/docker_rust_development
# Download the bash script
curl -L -s --output
# You can check the content of the bash script
# then run
# You will be asked to create a new passphrase for the SSh key. 
# Remember it, you will need it later.

That's it. Now we can test the connection from various locations.

1. Try the SSH connection from WSL2:

ssh -i ~/.ssh/rustdevuser_key -p 2201 rustdevuser@localhost
# Choose `yes` to save fingerprint if asked, just the first time.
# type passphrase
# should work !
# try for example
# ls result: rustprojects
# to exit the container

2. Try the SSH connection from windows command prompt or windows powershell terminal:

# test the ssh connection from Windows cmd prompt
"C:\WINDOWS\System32\OpenSSH\ssh.exe" -i ~\.ssh\rustdevuser_key -p 2201 rustdevuser@localhost
# Choose `y` to save fingerprint if asked, just the first time.
# type passphrase
# should work !
# try for example
# ls result: rustprojects
# to exit the container

3. Open VSCode and install extension Remote - SSH.

4. Then in VSCode F1, type ssh and choose Remote-SSH: Connect to Host... and choose rust_dev_pod.
Choose Linux if asked, just the first time.
Type your passphrase.
If we are lucky, everything works and you are now inside the container over SSH.

5. In VSCode terminal create a simple Rust project and run it:

cd ~/rustprojects
cargo new rust_dev_hello
cd rust_dev_hello
cargo run

That should work and greet you with "Hello, world!"

6. After reboot, WSL2 can create some network problems for podman.
We can simulate the WSL2 reboot in powershell in windows:

Get-Service LxssManager | Restart-Service

Before entering any podman command we need first to clean some temporary files, restart the pod and restart the SSh server.
In WSL2 terminal restart the pod after reboot:

sh ~/rustprojects/docker_rust_development/

7. Eventually you will want to remove the entire pod. Docker containers and pods are ephemeral, it means just temporary. But your code and data must persist. Before destroying the pod/containers, push your changes to github, because it will destroy also all the data that is inside.
Be careful !
In WSL2 terminal:

podman pod rm rust_dev_pod_create -f


Rust is a fantastic young language that empowers everyone to build reliable and efficient software. It enables simultaneously low-level control without giving up high-level conveniences. But with great power comes great responsibility !

Rust programs can do any "potentially harmful" things to your system. That is true for any compiled program, Rust is no exception.

But Rust can do anything also in the compile-time using and procedural macros. Even worse, if you open a code editor (like VSCode) with auto-completion (like Rust-analyzer), it will compile the code in the background without you knowing it. And the and procedural macros will run and they can do "anything" !

Even if you are very careful and avoid and procedural macros, your Rust project will have a lot of crates in the dependency tree. Any of those can surprise you with some "malware". This is called a "supply chain attack".

It is very hard to avoid "supply chain attacks" in Rust as things are today. We are just lucky, that the ecosystem is young and small and the malevolent players are waiting for Rust to become more popular. Then they will strike and strike hard. We need to be skeptical about anything that comes from the internet. We need to isolate/sandbox it so it cannot harm our system.

For big open-source project you will not read and understand every line of code. It is impossible because of the sheer size of projects and it is impossible to gain deep understanding of all the underlying principles, rules and exceptions of everything. And everything is moving and changing fast and continuously. It is impossible to follow all the changes.
We need to have layered protection between our computer system and some unknown code. In this project I propose a containerized development environment that will allow some degree of isolation. And in the same time easy of installation, transfer and repeatability.

Let learn to develope "everything" inside a Docker container and to isolate/sandbox it as much as possible from the underlying system.

I have to acknowledge that Docker Linux Containers are not the perfect sandboxing solution. But I believe that for my "Rust development environment", it is "good enough". I expect that container isolation will get better with time (google, amazon, Intel, OpenStack and IBM are working on it). My main system is Win10. Inside that, I have WSL2, which is a Linux virtual machine. And inside that, I have Docker Linux Containers.

Yes, there exists the possibility of abusing a kernel vulnerability, but I believe it is hard and they will get patched.
I didn't choose a true virtualization approach, but it is easy to run the container inside a virtual machine. More layers, more protection.

Docker and OCI

We all call it Docker, but Docker is just a well known company name with the funny whale logo. They developed and promoted Linux containers. Then they helped to create an open industry standards around container formats and runtimes. This standard is now called OCI (Open Container Initiative). So the correct name is "OCI containers" and "OCI images".

But "OCI" is a so boring name. I will still use the terms "Docker containers" and "Docker images" even if other companies can work interchangeably with them because of the OCI standards.

And there are alternatives to use Docker software that I will explore here.

Install Podman in Win10 + WSL2 + Debian 11(Bullseye)

Podman is a daemonless, open source, Linux native tool designed to work with Open Containers Initiative (OCI) Containers and Container Images. Containers under the control of Podman can be run by a non-privileged user. The CLI commands of the Podman "Container Engine" are practically identical to the Docker CLI. Podman relies on an OCI compliant "Container Runtime" (runc, crun, runv, etc) to interface with the operating system and create the running containers.

I already wrote some information how to install and use the combination of Win10 + WSL2 + Debian11(Bullseye):

Podman is available from the Debian11 package manager.

Let's install it. Open the WSL2 terminal and type:

sudo apt update
sudo apt install -y podman
podman version

Here we see some errors.

Wsl2 has a special kernel and Podman needs a small trick to work.

mkdir $HOME/.config/containers
nano $HOME/.config/containers/containers.conf

In this empty new file containers.conf write just this 3 lines and save:

cgroup_manager = "cgroupfs"
events_logger = "file"

Now you can command again and the result has no errors:

podman version
# 3.0.1

Sadly, the version on Debian 11 stable is 3.0.1, that is buggy and does not work properly. I will install the version 3.4.4 from the testing Debian 12. In WSL2 terminal:

sudo nano /etc/apt/sources.list
# add the line for `testing` repository
deb testing non-free contrib main
# Ctrl-o, enter, Ctrl-x
sudo apt update
# Then run 
sudo apt install -y podman
podman info
# podman 3.4.4
# Very important: Remove the temporary added line and run apt update
sudo nano /etc/apt/sources.list
# delete: deb testing non-free contrib main 
# Ctrl-o, enter, Ctrl-x
sudo apt update

Using Podman

Using the Podman CLI is just the same as using Docker CLI.

Inside the WSL2 terminal type:

podman images

We have no images for now. The words "image" and "container" are somewhat confusing. Super simplifying: When it runs it is a "container". If it does not run it is an "image". The image is just an "installation file". The containers can be started and stopped or attached and detached, but they are run only one time.

For test, run a sample container. It is a web server.

-d means it is run in a detached mode
-t adds a pseudo-tty to run arbitrary commands in an interactive shell
-p stays for publish port

The run command will download/pull the image if needed.

podman run -dt --name sample_cnt -p 8001:80/tcp

List all containers:

podman ps -a

Testing the httpd container:

curl http://localhost:8001

That should print the HTML page.

Finally you can remove the sample_cnt container we used:

podman rm sample_cnt

You can also remove the image, because this was just practice for learning:

podman rmi

Buildah for our Rust development images

Buildah is a replacement for the docker build command. It is easier to incorporate into scripts. It is pronounced exactly as Builder, but with a Bostonian accent ;-)

The Rust official images are on Docker hub:

I was surprised by the size of the image. It is big from 200Mb compressed to 1.2Gb uncompressed. But this is mostly the size of rust files.

I don't like that this images have only the root user. I will start from the Debian-11 image and install all I need as a non-privileged user rustdevuser.

In WSL2 terminal pull the image from Docker hub:

podman pull

I wrote the bash script

Run it with


This will create the image rust_dev_cargo_img.

The scripts sre just bash scripts and are super easy to read, follow, learn and modify. Much easier than Dockerfile. You can even run the commands one by one in the WSL2 terminal and inspect the container to debug the building process.

Rust development in a Docker container

There is a lot of benefits making a development environment in a container.

We want that everything is isolated/sandboxed and cannot affect our host system (WSL2 and Win10).

We also don't want to make any changes to our system because of Rust tools or our project needs.

We can have simultaneously more containers, each with a different version of Rust or a different toolchain with all the necessary configuration and tools. We can easily transfer the container to another system or to another developer and use it exactly as it is configured. Effortlessly.

We can save an container image with the source code and the exact state of all developer tools for a particular app version. Then years later we can still work on it for some security patches without the fear that new tools will break the old source code.

You will see that everybody uses podman run, but this is essentially 4 commands in one: pull from a repository, create the container, start or attach to the container and exec the bash in interactive mode. I like to use this commands separately, because it makes more sense for learning.

Create the container with a fixed name rust_dev_cargo_cnt:

--name - the container name will be rust_dev_cargo_cnt
-ti - we will use the container interactively in the terminal

podman create -ti --name rust_dev_cargo_cnt

We can list the existing containers with:

podman ps -a

Now we can start the container:

podman start rust_dev_cargo_cnt

Open the bash to interact with the container terminal:
-it - interactive terminal

podman exec -it rust_dev_cargo_cnt bash

We are now inside the container terminal and we can use cargo, rustup and other rust tools. The files we create will be inside the container. We are rustdevuser inside this container, so we will put our rustprojects in the /home/rustdevuser/rustprojects directory.

This container is started from Podman without root access to the host system !

This is a small, but important difference between Docker and Podman.

First let find the rustc version:

rustc --version
  rustc 1.61.0 (7737e0b5c 2022-04-04)

Let create and run a small Rust program:

cd ~/rustprojects
cargo new rust_dev_hello
cd rust_dev_hello
cargo run

That should work fine and greet you with "Hello, world!"

We can exit the container now with the command


When we exited the container we returned to WSL2 terminal.

The container still exists and is still running. Check with podman ps -a.

To interact with it again, repeat the previous command podman exec -it rust_dev_cargo_cnt bash.

This container does not work with VSCode and we will not need it any more. If you use another editor, you can use this image/container as base for your image/container for your editor.

Remove the container with:

podman rm rust_dev_cargo_cnt -f

Docker image with VSCode server and extensions

I use VSCode as my primary code editor in Windows. It works fine remotely with WSL2 (Debian Linux) with the Remote - WSL extension.

There exists a Remote - Containers extension, that lets you use a Docker container as a full-featured development environment. I tried it. There are a lot of automagic functions in this extension. But for some reason automagic never worked for me. I must always go painfully step-by-step and discover why the automagic does not work in my case. Always!

I decided to go with the SSH communication for remote developing. That is more broadly usable. We need to create an image that contains the VSCode server and extensions.

In WSL2 terminal run the bash script will create the image rust_dev_vscode_img.

Run it with:


This is based on the image rust_dev_cargo_img and adds the VSCode server and extensions. VSCode is great because of its extensions. Most of these extensions are installed inside the image rust_dev_vscode_img:

Other extensions you can add manually through VSCode, but then it is not repeatable. Better is to modify the script and recreate the image

Push image to docker hub

I signed in to

In Account settings - Security I created an access token. This is the password for podman login. It is needed only once.

Then I created a new image repository with the name rust_dev_vscode_img and tag latest. Docker is helping with the push command syntax. I use podman, so I just renamed docker to podman. The same for rust_dev_squid_img.

In WSL2 terminal:

podman login --username bestiadev
# type docker access token

podman push
podman push

podman push
podman push
podman push

podman push
podman push

It takes some time to upload more than 2 Gb.

enter the container as root

Sometimes you need to do something as root.
You don't need to use sudo. Just open the container as root user.

podman exec -it --user root rust_dev_vscode_cnt bash


Rust is not so small. I saved some 600MB of space just deleting the docs folder, that actually noone needs. squid3.5.27-2 168 MB cargo-1.61.0 1.11 GB cargo-1.61.0 1.40 GB

Users keys for SSH

We need to create 2 SSH keys, one for the SSH server identity host key of the container and the other for the identity of rustdevuser. This is done only once. To avoid old cryptographic algorithms I will force the new ed25519.
In WSL2 terminal:

# generate user key
ssh-keygen -f ~/.ssh/rustdevuser_key -t ed25519 -C "info@my.domain"
# give it a passphrase and remember it, you will need it
# generate host key
mkdir  -p ~/.ssh/rust_dev_pod_keys/etc/ssh
ssh-keygen -A -f ~/.ssh/rust_dev_pod_keys

# check the new files
# list user keys
ls -l ~/.ssh | grep "rustdevuser"
# -rw------- 1 luciano luciano  2655 Apr  3 12:03 rustdevuser_key
# -rw-r--r-- 1 luciano luciano   569 Apr  3 12:03

# list host keys
ls -l ~/.ssh/rust_dev_pod_keys/etc/ssh
# -rw------- 1 luciano luciano 1381 Apr  4 10:44 ssh_host_dsa_key
# -rw-r--r-- 1 luciano luciano  603 Apr  4 10:44
# -rw------- 1 luciano luciano  505 Apr  4 10:44 ssh_host_ecdsa_key
# -rw-r--r-- 1 luciano luciano  175 Apr  4 10:44
# -rw------- 1 luciano luciano  399 Apr  4 10:44 ssh_host_ed25519_key
# -rw-r--r-- 1 luciano luciano   95 Apr  4 10:44
# -rw------- 1 luciano luciano 2602 Apr  4 10:44 ssh_host_rsa_key
# -rw-r--r-- 1 luciano luciano  567 Apr  4 10:44

The same keys we will need in Windows, because the VSCode client works in windows. We will copy them: In WSL2 terminal:

echo $USERPROFILE/.ssh/rustdevuser_key
cp -v ~/.ssh/rustdevuser_key $USERPROFILE/.ssh/rustdevuser_key
cp -v ~/.ssh/ $USERPROFILE/.ssh/
cp -v -r ~/.ssh/rust_dev_pod_keys $USERPROFILE/.ssh/rust_dev_pod_keys
# check
ls -l $USERPROFILE/.ssh | grep "rustdevuser"
ls -l $USERPROFILE/.ssh/rust_dev_pod_keys/etc/ssh

Volumes or mount restrictions

I don't want that the container can access any file on my local system.

This is a "standalone" development container and everything must run inside.

The files can be cloned/pulled from Github or copied manually with podman cp.

Network Inbound restrictions

I would like to restrict the use of the network from/to the container.

When using Podman as a rootless user, the network is setup automatically. Only the localhost can be used. The container itself does not have an IP Address, because without root privileges, network association is not allowed. Port publishing as rootless containers can be done for "high ports" only. All ports below 1024 are privileged and cannot be used for publishing.

I hope/think that all inbound ports are closed by default and I need to explicitly expose them manually.

Network Outbound restrictions with Squid proxy in container

I would like to limit the access to the internet only to whitelisted domains:,,...

Some malware could want to "call home" and I will try to disable this.

What I need is a "proxy" or "transparent proxy". I will use the leading open-source proxy Squid, but in a container.

It can restrict both HTTP and HTTPS outbound traffic to a given set of Internet domains, while being fully transparent for instances in the private subnet.

I want to use this proxy for the container rust_dev_vscode_cnt. Container-to-container networking can be complex.

Podman works with pods, that make networking easy. This is usually the simplest approach for two rootless containers to communicate. Putting them in a pod allows them to communicate directly over localhost.

First create a modified image for Squid with (run inside the rustprojects/docker_rust_development folder):


If you need, you can modify the file etc_squid_squid.conf to add more whitelisted domains. Then run sh to build the modified image.

One pod with 2 containers

Podman and Kubernetes have the concept of pods, where more containers are tightly coupled. Here we will have the rust_dev_vscode_cnt that will use rust_dev_squid_cnt as a proxy. From outside the pod is like one entity with one address. All the network communication goes over the pod. Inside the pod everything is in the localhost address. That makes it easy to configure.

Inside the container rust_dev_vscode_cnt I want that everything goes through the proxy. This env variables should do that:

http_proxy, https_proxy,all_proxy.

Run the bash script to create a new pod rust_dev_pod with proxy settings:


The pod is now running:

podman pod list
podman ps -a 

SSH from WSL2 Debian

Try the SSH connection from WSL2 to the container:

ssh -i ~/.ssh/rustdevuser_key -p 2201 rustdevuser@localhost
# Choose `yes` to save fingerprint if asked, just the first time.
# type passphrase
# should work !
# try for example
# ls result: rustprojects
# to exit the container

SSH from windows

Run in windows cmd prompt to access the container over SSH from windows:

# test the ssh connection from Windows cmd prompt
"C:\WINDOWS\System32\OpenSSH\ssh.exe" -i ~\.ssh\rustdevuser_key -p 2201 rustdevuser@localhost
# Choose `y` to save fingerprint if asked, just the first time.
# type passphrase
# should work !
# try for example
# ls result: rustprojects
# to exit the container

debug SSH connection

Sometimes it is needed to debug the connection to the ssh server, because the normal error messages are completely useless.
From WSL2 terminal I enter the container terminal as root:

podman exec -it --user=root  rust_dev_vscode_cnt bash

In container terminal:

service ssh stop
/usr/sbin/sshd -ddd -p 2201
# now we can see the verbose log when we attach an SSH client to this server. And we can see where is the problem.
# after debug, start the service, before exit
service ssh start

To see the verbose log of the SSH client add -v like this:

ssh -i ~/.ssh/githubssh1 -p 2201 rustdevuser@localhost -v

To see the listening ports:

netstat -tan 


Open VSCode and install the extension Remote - SSH. In VSCode F1, type ssh and choose Remote-SSH: Open SSh configuration File..., choose c:\users\user_name\ssh\config and type (if is missing):

Host rust_dev_pod
  HostName localhost
  Port 2201
  User rustdevuser
  IdentityFile ~\.ssh\rustdevuser_key
  IdentitiesOnly yes

Save and close.

Then in VSCode F1, type ssh and choose Remote-SSH: Connect to Host... and choose rust_dev_pod.
Choose Linux if asked, just the first time.
Type your passphrase.
If we are lucky, everything works and VSCode is now inside the container over SSH.

VSCode terminal

VSCode has an integrated VSCode terminal. It has some advantages for developers that the standard WSL2 terminal does not have. It is great to use it for everything around code in containers. You can open more then one VSCode terminal if you need to. For example if you run a web server.

If the VSCode terminal is not opened simply press Ctrl+j to open it and the same to close it.

Inside VSCode terminal, we will create a sample project:

cd ~/rustprojects
cargo new rust_dev_hello
cd ~/rustprojects/rust_dev_hello

This easy command opens a new VSCode window exactly for this project/folder inside the container:

code .

A new VSCode windows will open for the rust_dev_hello project. Because of the SSH communication it asks for the passphrase. You can close now all other VSCode windows.

Build and run the project in the VSCode terminal:

cargo run

That should work and greet you with "Hello, world!".

Leave VSCode open because the next chapter will continue from here.

Github in the container

Shortcut: Download the template for bash script from here: into Win10 folder ~\.ssh. It contains all the steps from below. You need to personalize it.
Then copy it manually into WSL2 - just once. Run in WSL Terminal:

cp -v $USERPROFILE/.ssh/ ~/.ssh/
# and then run
sh ~/.ssh/

Manually step by step instructions that are inside the
Git inside the container does not yet have your information, that it needs: In WSL2 terminal:

podman exec --user=rustdevuser rust_dev_vscode_cnt git config --global "info@your.mail"
podman exec --user=rustdevuser rust_dev_vscode_cnt git config --global "your_name"
podman exec --user=rustdevuser rust_dev_vscode_cnt git config --global -l

I like to work with Github over SSH and not over https. I think it is the natural thing for Linux.
To make SSH client work in the container I need the file with the private key for SSH connection to Github. I already have this in the file ~/.ssh/githubssh1. I will copy it into the container with podman cp.
Be careful ! This is a secret ! It means that this container I cannot share anymore with anybody. It is now my private container. I must never make an image from it and share it. Never !

In WSL2 terminal:

podman exec --user=rustdevuser rust_dev_vscode_cnt ls -l /home/rustdevuser/.ssh
podman cp ~/.ssh/githubssh1 rust_dev_vscode_cnt:/home/rustdevuser/.ssh/githubssh1
podman exec --user=rustdevuser rust_dev_vscode_cnt chmod 600 /home/rustdevuser/.ssh/githubssh1
podman cp ~/.ssh/ rust_dev_vscode_cnt:/home/rustdevuser/.ssh/
podman exec --user=rustdevuser rust_dev_vscode_cnt ls -l /home/rustdevuser/.ssh

The VSCode terminal is still opened on the project rust_dev_hello.

SSH Agent

It is comfortable to use the ssh-agent to store the passphrase in memory, so we type it only once. The ssh-agent is already started on login in the ~/.bash_profile script.

Again attention, that this container has secrets and must not be shared ! Never !

In the VSCode terminal (Ctrl+j) run:

ssh-add /home/rustdevuser/.ssh/githubssh1
# enter your passphrase

You can download the template from Github and save it into win10 folder ~\.ssh. There, personalize it with you SSH keys file names. Then copy it manually into WSL2 and the container - just once. Run in WSL Terminal:

cp -v $USERPROFILE/.ssh/ ~/.ssh/  
podman cp ~/.ssh/ rust_dev_vscode_cnt:/home/rustdevuser/.ssh/  

You will then run inside the VSCode terminal for each window/project separately:

sh ~/.ssh/

After you enter the passphrase, it will remember it until the terminal is open.
When you open the terminal again, you will have to run the script again and enter the passphrase again.

Github push

Open in the browser and sign in, click New and create a new repository named rust_dev_hello.

Github is user friendly and shows the standard commands we need to run.

Modify the commands below to your Github repository.

In VSCode click on the Source control and click Initialize and type the commit msg init. Then in VSCode terminal run:

git remote add origin
git push -u origin main

Done! Check your Github repository. Always push the changes to Github. So you can destroy this pod/container and create a new empty one, then pull the code from Github and continue developing. Containers are the worst place to have persistent data stored. They can be deleted any second for any reason. Leave VSCode open because the next chapter will continue from here.

Existing Rust projects on Github

You probably already have a Rust project on Github. You want to continue its development inside the container.

For an example we will use my PWA+WebAssembly/WASM project rust_wasm_pwa_minimal_clock, that needs to forward the port 8001, because our project needs a web server. That is fairly common. I am not a fan of autoForward automagic in VSCode, so I disable it in File-Preferences-Settings search remote.autoForwardPorts and uncheck it to false.

We will continue to use the existing VSCode terminal, that is already opened on the folder /home/rustdevuser/rustprojects/rust_dev_hello. Just to practice.

Run the commands to clone the repository from Github and open a new VSCode window. We already have the SSH private key and ssh-agent running:

cd /home/rustdevuser/rustprojects/
git clone
cd rust_wasm_pwa_minimal_clock
code .

The code . command will open a new VSCode window in the folder rust_wasm_pwa_minimal_clock. Enter the SSH passphrase if asked. In the new VSCode window, we can now edit, compile and run the project. All sandboxed/isolated inside the container. We can now close the other VSCode windows, we don't need it any more.

This example is somewhat more complex, because it is WebAssembly, but it is good for learning. I used the automation tool cargo-auto to script a more complex building process. You can read the automation task code in automation_task_rs/src/ On the first build it will download the wasm components and wasm-bindgen. That can take some time. Don't despair.

Now we can build and run the project in VSCode terminal (Ctrl+j):

cargo auto build_and_run

In VSCode go to Ports and add the port 4000.

Open the browser in Windows:


This is an example of Webassembly and PWA, directly from a docker container.

A good learning example.

After reboot

After reboot WSL2 can create some network problems for podman.
We can simulate the WSL2 reboot in powershell in windows:

Get-Service LxssManager | Restart-Service

Before entering any podman command we need first to clean some temporary files, restart the pod and restart the SSh server.
In WSL2 terminal:


VSCode and file copying from win10

It is easy to copy files from Win10 to the VSCode project within the container just with drag&drop.

Protect the SSH private key in Windows

In Linux the private keys inside ~/.ssh are protected with chmod 600. We need to do similarly for the private keys inside Win10 folder ~\.ssh Run in powershell terminal with the standard user:

cd ~/.ssh
# Set Key File Variable:
  New-Variable -Name Key -Value "$env:UserProfile\.ssh\id_rsa"
# Remove Inheritance:
  Icacls $Key /c /t /Inheritance:d
# Set Ownership to Owner:
  Icacls $Key /c /t /grant ${env:UserName}:F
# Remove All Users, except for Owner:
  Icacls $Key  /c /t /Remove Administrator BUILTIN\Administrators BUILTIN Everyone System Users
# Verify:
  Icacls $Key
# Remove Variable:
  Remove-Variable -Name Key


It is a complex setting. There can be some quirks sometimes.

ssh could not resolve hostname

Warning: The "ssh could not resolve hostname" is a common error. It is not that big of an issue. I closed everything and restart my computer and everything works fine now.


SSH client remembers the key of the servers in the file ~/.ssh/known_hosts.
If we created a new key for the ephemeral container, we can get the error REMOTE HOST IDENTIFICATION HAS CHANGED. It is enough to open the file ~/.ssh/known_hosts and delete the offending line. In WSL2 terminal we can use:

ssh-keygen -f ~/.ssh/known_hosts -R "[localhost]:2201";
ssh-keygen -f $USERPROFILE/.ssh/known_hosts -R "[localhost]:2201";

Typescript compiler

Some projects need the typescript compiler tsc. First we need to install nodejs with npm to install typescript. That is a lot of installation. This is because I don't want it in my default container. For typescript I created a new container image: rust_ts_dev_image.
The bash script sh will create the new image with typescript.
Then we can use sh to create the podman pod with typescript.
And there is sh to use it after reboot.
You cannot use the pods rust_dev_pod and rust_ts_dev_pod simultaneously. You must run one or the other, because they use the same ports.

PostgreSQL and pgAdmin

Some projects need the database PostgreSQL 14. I created a new pod rust_pg_dev_pod with the command sh You cannot use the pods rust_dev_pod and rust_pg_dev_pod simultaneously. You must run one or the other, because they use the same ports.
And there is sh to use it after reboot.
To use the administrative tool pgAdmin open the browser on localhost:9876.
You can change the user and passwords in the bash script to something stronger.

Read more

Read more how I use my Development environment.


Watch the log if the access is restricted to some domains: podman exec rust_dev_squid_cnt cat /var/log/squid/access.log podman exec rust_dev_squid_cnt tail -f /var/log/squid/access.log

new image with cargo-crev and cargo_crev_reviews new image with windows cross compile

Add this following to your ~/.bash_profile


cargo crev reviews and advisory

We live in times of danger with supply chain attacks.
It is recommended to always use cargo-crev
to verify the trustworthiness of each of your dependencies.
Please, spread this info.
You can also read reviews quickly on the web:

open-source and free as a beer

My open-source projects are free as a beer (MIT license).
I just love programming.
But I need also to drink. If you find my projects and tutorials helpful,
please buy me a beer donating on my paypal.
You know the price of a beer in your local bar ;-)
So I can drink a free beer for your health :-)
Na zdravje! Alla salute! Prost! Nazdravlje! 🍻