You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am developing a GPU accelerated IoT Edge module on an Nvidia Jetson Xavier device that requires setting the Docker runtime to "nvidia" to ensure that created modules are bootstrapped with nvidia-docker to allow appropriate access to the system GPU resources and CUDA librarires.
The module settings from my deployment.template.json can be seen below:
Hi, @toolboc , from our investigation, only docker-compose.yaml with version 2.3, 2.4 supported runtime config, and it is not supported by version 3.x.
We are looking to clean up the reported issues that are obsolete such that our team can focus on the latest issues reported.
As such, we are going to close the issue for now.
If the issue is still blocking or critical to you, please file a new issue and we will do our best to respond as soon as possible.
When you file a new issue, please ensure following additional information are provided.
o If you believe it’s a bug, provide detail repro steps.
o If it’s a feature request, please note as such
o Expected outcome
o Is this a blocking issue?
I am developing a GPU accelerated IoT Edge module on an Nvidia Jetson Xavier device that requires setting the Docker runtime to "nvidia" to ensure that created modules are bootstrapped with nvidia-docker to allow appropriate access to the system GPU resources and CUDA librarires.
The module settings from my deployment.template.json can be seen below:
These settings work perfectly fine in a deployment configuration and set the Docker Runtime appropriately:
However, when the DeepStreamModule is created using the IoT Edge Simulator, I get the following result:
Is it possible to override this behavior to allow setting of the Docker Runtime to something other than "runc"?
The text was updated successfully, but these errors were encountered: