Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Drop Ubuntu 20.04 support for apt dependencies #1439

Open
2 of 3 tasks
traversaro opened this issue Jul 6, 2023 · 7 comments
Open
2 of 3 tasks

Drop Ubuntu 20.04 support for apt dependencies #1439

traversaro opened this issue Jul 6, 2023 · 7 comments

Comments

@traversaro
Copy link
Member

traversaro commented Jul 6, 2023

I know that we did not fully even solved #481. Anyhow the problems related to Ubuntu 20.04 apt dependencies are increasting, so I wanted to start opening this issue to start collecting things that block us from updating to 20.04:

  • JetPack SDK stuck to Ubuntu 20.04
  • iCubOS not running on Ubuntu 20.04
  • AMI machine(s) not updated to Ubuntu 22.04

@Nicogene do you know the situation of icub-tech's machines and Docker images?
@xEnVrE do you know about the situation of HSP's machine or other constraints (like in #481 we had CUDA-related problems).

@xEnVrE
Copy link
Collaborator

xEnVrE commented Jul 6, 2023

Hi @traversaro,

at the moment all the machines (personal and those of iCubGenova11) are mostly using Ubuntu 20.04 - but I counted at least 3 machines already using Ubuntu 22.04.

I would say we would need to update at least those of iCubGenova11 when the support is removed. I cannot comment on the machines of the R1 setup - not sure they are using the superbuild.

As regards applications w/ CUDA they mostly run on docker where we have, I think, a mix of situations (ranging from Ubuntu 18.04 to Ubuntu 22.04) and not all of them require the software provided by the superbuild.

@Nicogene
Copy link
Member

Nicogene commented Jul 6, 2023

@Nicogene do you know the situation of icub-tech's machines and Docker images?

@valegagge can confirm it but I think that the docker images are focal-based:

https://github.com/icub-tech-iit/docker-deployment-images/blob/2961c08bc4fc39c75e03adbf2d5371fddcb2723b/dockerfile_images/basic/superbuild/conf_build.ini#L2

@valegagge
Copy link
Member

Hi @Nicogene ,

@valegagge can confirm it but I think that the docker images are focal-based:

yes!I confirm.
When we are ready to move to the 22 we can start to update the docker images and our laptops.

@traversaro
Copy link
Member Author

Possibly related PR: robotology/yarp#2880 .

@traversaro
Copy link
Member Author

Since robotology/yarp#3048 YARP does not officially support anymore Ubuntu 20.04 with apt dependencies, and after robotology/yarp#3155 YARP master is not compiling anymore with Ubuntu 20.04 with apt dependencies. Unless someone wants to contribute some work on fixing the compilation of YARP on Ubuntu 20.04 with apt dependencies, I think we should just drop support for it.

For what regards iCubOS, we had icub-tech-iit/documentation#381 while there is still icub-tech-iit/documentation#383, so hopefully that will be fixed soon.

For what regards:

AMI machine(s) not updated to Ubuntu 22.04

I will check if this still a problem, but if that is the case we can just use conda there.

JetPack SDK stuck to Ubuntu 20.04

Again, also here I do not think there are a lot of alternatives to use conda or docker. @Nicogene @martinaxgloria do you think there is any docs on this to update?

@Nicogene
Copy link
Member

Again, also here I do not think there are a lot of alternatives to use conda or docker. @Nicogene @martinaxgloria do you think there is any docs on this to update?

I don't think there is any doc on this, this is quite a big issue right now both because we still have some Xavier around that are stuck on jetpack 5 (ubuntu 20.04 based), and we found that jetpack 6 has some problems with the display we are using on ergocub, so it is problematic also for the Orin we have.

For the time being, I think we have to write some documentation for compiling the superbuild on those boards with conda dependencies

@traversaro
Copy link
Member Author

For the time being, I think we have to write some documentation for compiling the superbuild on those boards with conda dependencies

The nice news on that is that we now have up-to-date humble and jazzy packages in robostack, so one limitation that prevented us to use conda there is not present anymore.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants