Platform advice - what do I do if I'm on ARM?

[johnwee1]

Hi, I have a macbook with apple silicon and want to follow this tutorial.

At first I tried to replicate the setup by following the deps in the dockerfile but noticed a few are missing in robostack conda forge repo, namely anything andino, ros-humble-turtle-nest, the transformation lib etc...

After searching, I find that it's possible to use UTM to create a ubuntu VM, but using a VM would still mean the architecture is arm (and therefore not officially covered by this tutorial).

There is an option to run UTM as an emulator instead which would allow me to emulate a x64 ubuntu machine, but I was wondering if that was worth doing since there is a performance penalty. Does anyone know if this penalty is going to be large (i.e. should I not bother with ARM and just use a spare old windows laptop lying around?)

[ilkkaj]

Hi, one of my students in the course used UTM in Apple Silicon (not sure which chip it was, perhaps M2 on MacBook Air), and he was able to run Ubuntu and the Docker container successfully. I will investigate this more and come back to you if I find any further info how he did it. We are definitely trying to make the ARM version to work, there would be a clear need for it.

[alexandroid000]

I have a similar issue but it's that students do not have permission to use Docker on campus computers. We worked around this by implementing an Apptainer (all our Linux machines are on Debian). However, I now have the same issue of installing dependencies and exercises. Are the packages in the packages directory all we need, or do you have a stand-alone ROS2 package for the course?

Thank you very much for this resource, kiitos!

[Jannkar]

I have a similar issue but it's that students do not have permission to use Docker on campus computers. We worked around this by implementing an Apptainer (all our Linux machines are on Debian). However, I now have the same issue of installing dependencies and exercises. Are the packages in the packages directory all we need, or do you have a stand-alone ROS2 package for the course?

Thank you very much for this resource, kiitos!

Thanks for the question! If you are looking to manually install all the dependencies, the best way is to take a look at the Dockerfile, and go through every step. Making a stand-alone ROS 2 package of the course is an interesting idea, but we don't have that at the moment. The course depends on many packages, and some of them need to be installed manually (such as gazebo, nav2, turtle-nest, andino), and some of them can be installed using rosdep. Also, the packages directory includes some custom dependencies.

[8gaU8]

Hi! I am the student who was using a MacBook Pro with Apple Silicon M1 during the class. I was using UTM, but I needed to customize several configurations. What I did are:

• Operating System: Ubuntu 22.04.5 LTS
• Architexture: aarch64
• System: QEMU virt-7.2
• Display card: virtio-gpu-pci ( NO GL )

It will take time to boot, but it worked for me.
Hope it will help

[willnearn]

I have a similar issue but it's that students do not have permission to use Docker on campus computers. We worked around this by implementing an Apptainer (all our Linux machines are on Debian). However, I now have the same issue of installing dependencies and exercises. Are the packages in the packages directory all we need, or do you have a stand-alone ROS2 package for the course?

Thank you very much for this resource, kiitos!

Hi Alex @alexandroid000, if you're running Debian 11 Bullseye or newer, you should have podman, which can make containers based on Dockerfiles. A cursory search shows that Apptainer can't do this (sounds like your issue here?), but I'm not very familiar with Apptainer. Unless your admins have podman blocked as well as Docker, you should probably be good to switch directly to podman :)

Let me know how this works for ya🤙🤙