Hi,
I’m using the Dockerfile.l4t-humble file to create a docker container on an NVIDIA Jetson AGX Orin. However I’m hitting connection errors on the RUN command with the jp6/cu126/protobuf/. These are provided in the attached file (which contains the last part of my terminal output where the errors showed up)
error_log.txt (5.1 KB)
. Any tips? I’m on Ubuntu 22.04, ROS2 Humble, R36.4.4, trying to run l4t-r36.4.0, zedsdk-5.0.3. Thanks
Have you followed the documentation to add a new user?
It seems the issue is coming from “pip install protobuf”. To confirm this, can you comment out this part in the RUN command and retry? If this is the issue, it doesn’t concern the ZED SDK.
Protobuf is kinda notorious for being tricky with dependencies and versioning, make sure you absolutely need it
Yep I can try commenting it out. However that’s come from the l4t-humble dockerfile not something I added in myself, so I’d been thinking it was necessary for ZED SDK?
Ok, I thought this was a custom Docker image.
protobuf is not needed by the ZED SDK, but something in this image might need it. The issue is coming from an nvidia server being down/redirected see this Github issue: PyPI index (http://jetson.webredirect.org) is down · Issue #811 · dusty-nv/jetson-containers · GitHub
I’ll check how to make a permanent fix
1 Like
All good thanks for looking into it. If it helps, I was curious and asked an ai and it replaced the pip install protobuf with pip install protobuf --index-url https://pypi.org/simple. I was then able to launch and enter the container. I’m relatively new to this field so I don’t fully understand/trust the fix but thought I should pass it on if it’s useful.
But I also haven’t been able to properly verify it since the zed packages aren’t building in the container, but that’s a different jetson/cuda related issue I’m working on.
1 Like