Docker is a popular solution designed to contain a single application: for example, an MySQL database server. LXD is similar in some ways but designed to contain an entire operating system. This makes it useful for some scenarios. For example, you can spin up an LXD container, install a database server and an http server. You can then create a WordPress website inside. You can now switch from cloud to cloud by just moving this LXD container where it’s needed when you’re not happy with your previous provider. And since it’s easy to clone a container, you can even upload your website to multiple cloud providers to create a redundant, high-availability setup. LXD doesn’t virtualize hardware like QEMU or VirtualBox do, which means it’s very fast, offering near-native speed of execution.
Install and Configure LXD
Open a terminal and install LXD, plus the ZFS utilities, which will help you speed up some operations and save disk space when working with containers.
Start LXD configuration. Press ENTER to select the default values. For Size in GB of the new loop device (1GB minimum) [default=15GB], you can pick another value, like “50GB” if you know you’ll create a lot of containers.
Find and Launch an LXD Distribution Image
To list all Ubuntu images:
There will be a lot of results. You can ommit arch=amd64 if you need images for other platforms, like ARM processors (Raspberry Pi devices use such architecture). In the previous picture the results have been limited (with |head) to make it easier to read. The fingerprint of Ubuntu 18.04 (84a71299044b) has been highlighted. If you want to launch a container with that distribution, the command would be: At the moment this would return a permission denied error. You need to be in the “lxd” group. Your user has been added to this group already, but to make it active you have to log out and log back in. If you want to avoid that, use this command, replacing “user” with your username:
Now, lxc commands work without requiring sudo.
Launch LXD Containers with non-Ubuntu Distributions
This command will show you what other distributions are available: To launch an image, instead of the fingerprint, you can also use an alias name if you see one available in that list.
If you append a string at the end, you can choose a name for your container:
Manage LXD Containers
To list all containers:
The “IPV4” column especially is important if you have any running services on that instance. For example, if an Apache http server would be running on the instance, entering “10.234.232.246” in the browser would display the website hosted in the container. To stop a container: This can take a very long time (or fail) with non-Ubuntu distributions. It’s better to get a shell to the container, and once inside, enter systemctl poweroff to stop it. If all else fails, you can force a stop with: To start it: To move inside your container:
You can install programs with “sudo apt install” and do anything else you would do on a normal Linux distribution, e.g., configure an Apache server. When you want to exit from the container, simply type:
Transfer Files to/from LXD Containers
To upload a file to your container: Include the name of the file to be created, not only the directory where you want to upload it. Here is an example: To upload a directory instead of a file: Example: To download a directory from your container to your main operating system: Example:
Conclusion
This covers the basic usage of LXD containers. There are more advanced features such as snapshots and rollbacks, imposing limits on resources such as CPU and RAM, cloning containers, and so on. These may be covered in a future tutorial if we notice readers are interested in the subject.
