- For Docker images: Change the base image of the Runtime Container to the latest supported Ubuntu release.
- Create a MATLAB docker container based on the correct base image to perform the compilation inside that container (https://www.mathworks.com/help/cloudcenter/ug/matlab-container-on-docker-hub.html )
Why do I receive errors like "/lib/x86_64-linux-gnu/libc.so.6: version 'GLIBC_2.34' not found" when running Docker images or WebApps that contain Simulink simulations in certain situations?
23 views (last 30 days)
Show older comments
MathWorks Support Team
on 11 Jun 2024
Answered: MathWorks Support Team
on 11 Jun 2024
Why do I receive errors like "/lib/x86_64-linux-gnu/libc.so.6: version 'GLIBC_2.34' not found" when running Docker images or WebApps that contain Simulink simulations in certain situations?
The path or the GLIBC version, that is not found, can be different.
This issue is not tied to a specific release.
Accepted Answer
MathWorks Support Team
on 12 Jun 2024
The root cause of this issue is that the application is compiled on another (newer) OS with a different GLIBC version than the (older) OS that is running the application.
The reason for these errors to occur is the backward compatibility of GLIBC. New versions of GLIBC contain new symbols, but they do not always have new names for the linked libraries.
The given applications of Docker containers and WebApps are candidates where the application can easily run on another operating system than what was used to compile the application.
For docker containers, the MATLAB Runtime Baseimage is based on an Ubuntu image (the exact version used is dependent on the MATLAB release), but it is not guaranteed that this is the latest supported Ubuntu version, so this issue can occur when a newer version of Ubuntu is used to compile the model.
Additionally, similar errors can also occur when the compilation is done on Ubuntu/Debian-based systems whereas deployment is on RHEL/SUSE-based systems or vis-versa.
If you run into these issues, try to align the systems in a way that compilation and deployment are done on the same family of distributions and that the deployment system is the same or newer than the compilation system.
To achieve that, you can avoid having to set up a completely new operating system by trying one of the following workflows:
0 Comments
More Answers (0)
See Also
Categories
Find more on Containers in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!