Przejdź do treści


There are several ways to run a copy of mkdocs so that you can see exactly how your Rocky Linux document will appear when it is merged on the live system. This particular document deals with using an LXD container on your local workstation to separate python code in mkdocs from other projects you might be working on.

It is recommended to keep projects separate to avoid causing problems with your workstation's code.

This is also a partner document to the Docker version here.

Prerequisites and Assumptions

A few things you should have/know/be:

  • Familiarity and comfort with the command-line
  • Comfortable using tools for editing, SSH, and synchronization, or willing to follow along and learn
  • We will reference LXD - there's a long document on building and using LXD on a server here, but we will be using just a basic install on our Linux workstation. This document assumes that you are already using LXD for other things, and does not cover the build and initialization of LXD.
  • We will be using lsyncd for mirroring files, and you can find documentation on that here
  • You will need public keys generated for your user and the "root" user on your local workstation using this document
  • Our bridge interface is running on and our container is running on in our examples below.
  • "youruser" in this document represents your user id, so substitute in your own.
  • We are assuming that you are already doing documentation development with a clone of the documentation repository on your workstation.

The mkdocs container

Create the Container

Our first step is to create the LXD container. There's no need to use anything other than defaults here, so allow your container to be built using the bridge interface.

We will add a Rocky container to our workstation for mkdocs, so we are just calling it "mkdocs":

lxc launch images:rockylinux/8 mkdocs

The container needs to be set up with a proxy. By default, when mkdocs serve is started, it runs on That's fine when it is on your local workstation without a container. However, when it is in an LXD container on your local workstation, you need to set up the container with a proxy port. This is done with:

lxc config device add mkdocs mkdocsport proxy listen=tcp: connect=tcp:

In the line above, "mkdocs" is our container name, "mkdocsport" is an arbitrary name we are giving to the proxy port, the type is "proxy", and then we are listening on all tcp interfaces on port 8000 and connecting to the localhost for that container on port 8000.


If you're running the lxd instance on another machine in your network, remember to make sure that port 8000 is open in the firewall.

Installing Packages

First, get into the container with:

lxc exec mkdocs bash

Changes in requirements.txt for 8.x

The current requirements.txt will require a newer version of Python than what is installed by default in Rocky Linux 8.5 or 8.6. To be able to install all the other dependencies, do the following:

sudo dnf module enable python38
sudo dnf install python38

You can then skip installing python3-pip in the packages found below.

We will need a few packages to accomplish what we need to do:

dnf install git openssh-server python3-pip rsync

Once installed, we need to enable and start sshd:

systemctl enable --now sshd

Container Users

We need to set a password for our root user, and then add our own user (the user that you use on your local machine) and add it to the sudoers list. We should be the "root" user at the moment, so to change the password type:


And set the password to something secure and memorable.

Next, add your user and set a password:

adduser youruser
passwd youruser

And add your user to the sudoers group:

usermod -aG wheel youruser

At this point, you should be able to SSH into the container using either the root user or your user from your workstation and entering a password. Make sure that you can do that before continuing.

SSH for root and Your user

In this procedure, the root user (at minimum) needs to be able to SSH into the container without entering a password; this is because of the lsyncd process we will be implementing. We are assuming here that you can sudo to the root user on your local workstation:

sudo -s

We are also assuming that the root user has an key in the ./ssh directory. If not, generate one using this procedure:

ls -al .ssh/
drwx------  2 root root 4096 Feb 25 08:06 .
drwx------ 14 root root 4096 Feb 25 08:10 ..
-rw-------  1 root root 2610 Feb 14  2021 id_rsa
-rw-r--r--  1 root root  572 Feb 14  2021
-rw-r--r--  1 root root  222 Feb 25 08:06 known_hosts

To get SSH access on our container without having to enter a password, as long as the key exists, like it does above, all we need to do is run:

ssh-copy-id root@

In the case of our user, however, we need the entire .ssh/ directory copied to our container. The reason is that we are going to be keeping everything identical for this user so that our access to GitHub over SSH is the same.

To copy everything over to our container, we just need to do this as your user, not sudo:

scp -r .ssh/ youruser@

Next, SSH into the container as your user:

ssh -l youruser

We need to make things identical, and this is done with ssh-add. To do this, though, we need to make sure that we have the ssh-agent available. Enter the following:

eval "$(ssh-agent)"

Cloning repositories

We need two repositories cloned, but there is no need to add any git remotes. The documentation repository here will only be used to display the current documentation (mirrored from your workstation) and the docs.

The repository will be used for running mkdocs serve and will use the mirror as it's source. All of these steps should be done as your non-root user. If you are not able to clone the repositories as your userid, then there IS a problem with your identity as far as git is concerned and you will need to review the last few steps for re-creating your key environment (above).

First, clone the documentation:

git clone

Assuming that worked, then clone the

git clone

If everything worked as planned, then you can move on.

Setting Up mkdocs

Installing the needed plugins is all done with pip3 and the "requirements.txt" file in the directory. While this process will argue with you about using the root user for this, in order to write the changes to the system directories, you pretty much have to run it as root.

We are doing this with sudo here.

Change into the directory:


Then run:

sudo pip3 install -r requirements.txt

Next we need to set up mkdocs with an additional directory. Right now, mkdocs requires a docs directory to be created and then the documentation/docs directory linked beneath it. All this is done with:

mkdir docs
cd docs
ln -s ../../documentation/docs

Testing mkdocs

Now that we have mkdocs setup, let's try starting the server. Remember, this process is going to argue that it looks like this is production. It's not, so ignore the warning. Start mkdocs serve with:

mkdocs serve -a

You should see something like this in the console:

INFO     -  Building documentation...
WARNING  -  Config value: 'dev_addr'. Warning: The use of the IP address '' suggests a production environment or the use of a
            proxy to connect to the MkDocs server. However, the MkDocs' server is intended for local development purposes only. Please
            use a third party production-ready server instead.
INFO     -  Adding 'sv' to the '' option
INFO     -  Adding 'it' to the '' option
INFO     -  Adding 'es' to the '' option
INFO     -  Adding 'ja' to the '' option
INFO     -  Adding 'fr' to the '' option
INFO     -  Adding 'pt' to the '' option
WARNING  -  Language 'zh' is not supported by lunr.js, not setting it in the '' option
INFO     -  Adding 'de' to the '' option
INFO     -  Building en documentation
INFO     -  Building de documentation
INFO     -  Building fr documentation
INFO     -  Building es documentation
INFO     -  Building it documentation
INFO     -  Building ja documentation
INFO     -  Building zh documentation
INFO     -  Building sv documentation
INFO     -  Building pt documentation
INFO     -  [14:12:56] Reloading browsers

Now for the moment of truth! If you've done everything correctly above, you should be able to open a web browser and go to the IP of your container on port :8000, and see the documentation site.

In our example, we would enter the following in the browser address (NOTE To avoid broken URLs, the IP here has been changed to "your-server-ip". You just need to substitute in the IP):



If you saw the documentation in the web browser, we are almost there. The last step is to keep the documentation that's in your container in synchronization with the one on your local workstation.

We are doing this here with lsyncd as noted above.

Depending on which Linux version you are using, lsyncd is installed differently. This document covers ways to install it on Rocky Linux, and also from source. If you are using some of the other Linux type (Ubuntu for example) they generally have their own packages, but there are nuances to them.

Ubuntu's, for example, names the configuration file differently. Just be aware that if you are using another Linux workstation type other than Rocky Linux, and don't want to install from source, there are probably packages available for your platform.

For now, we are assuming that you are using a Rocky Linux workstation and are using the RPM install method from the included document.



The root user must run the daemon, so you will need to be root to create the configuration files and logs. For this we are assuming sudo -s.

We need to have some log files available for lsyncd to write to:

touch /var/log/lsyncd-status.log
touch /var/log/lsyncd.log

We also need to have an exclude file created, even though in this case we are not excluding anything:

touch /etc/lsyncd.exclude

Finally we need to create the configuration file. In this example, we are using vi as our editor, but you may use whichever editor you feel comfortable with:

vi /etc/lsyncd.conf

And then place this content in that file and save it. Be sure to replace "youruser" with your actual user, and the IP address with your own container IP:

settings {
   logfile = "/var/log/lsyncd.log",
   statusFile = "/var/log/lsyncd-status.log",
   statusInterval = 20,
   maxProcesses = 1

sync {
   rsync = {
     archive = true,
     compress = false,
     whole_file = false
   ssh = {
     port = 22

We are assuming that you enabled lsyncd when you installed it, so at this point we need to just start or restart the process:

systemctl restart lsyncd

To make sure things are working, check the logs-particularly the lsyncd.log, which should show you something like this if everything started correctly:

Fri Feb 25 08:10:16 2022 Normal: --- Startup, daemonizing ---
Fri Feb 25 08:10:16 2022 Normal: recursive startup rsync: /home/youruser/documentation/ -> root@
Fri Feb 25 08:10:41 2022 Normal: Startup of "/home/youruser/documentation/" finished: 0
Fri Feb 25 08:15:14 2022 Normal: Calling rsync with filter-list of new/modified files/dirs


As you work on your workstation documentation now, whether it is a git pull or a branch you create to make a document (like this one!), you will see the changes appear in your documentation on the container, and mkdocs serve will show you the content in your web browser.

It's recommended practice that all Python code should be run separate from any other Python code you might be developing. LXD containers can make that a lot easier; give this method a try and see if it works for you.

Ostatnia aktualizacja: 3 marca 2023

Author: Steven Spencer

Contributors: Ezequiel Bruni