Newbie trying to get C++ template app working on Verdin/Dalhia

New to Toradex, pretty much new to Linux. Purchased Verdin & Dalhia boards, went through quick start. Then created new Torizon project using C++ template (hello world). I get an error during the push/pull of the image. Also, I have Docker Desktop installed (host is AMD Ryzen Win11).

Here is the last of the build terminal output:

Executing task: xonsh /home/compass/opencvtest/.conf/service-check.xsh registry compass 22 torizon 192.168.12.160 ‘curl --silent --max-time 5 http://localhost:5002/v2/_catalog’ Registry ready

* Terminal will be reused by tasks, press any key to close it.

* Executing task: sleep 1 && DOCKER_HOST=192.168.12.160:2375 docker pull localhost:5002/opencvtest-debug:arm64 Error response from daemon: manifest for localhost:5002/opencvtest-debug:arm64 not found: manifest unknown: manifest unknown

* The terminal process “/usr/bin/bash ‘-c’, ‘sleep 1 && DOCKER_HOST=192.168.12.160:2375 docker pull localhost:5002/opencvtest-debug:arm64’” terminated with exit code: 1.

Regarding Docker Desktop, when I go to run it most times it doesn’t display the app window but can be found in Task Manager under Background Processes. One must run the Docker Desktop before building as DD starts the docker daemon that the Torizon IDE build requires.

Hi, I’m the Toradex AI Assistant.

I might make mistakes or occasionally get you wrong. Our support team will step in if needed.


I’d be glad to help! Could you let me know which OS and version you are using?

To better assist you with this container build issue, I need to understand your specific setup. Please run the tdx-info command on your Verdin module to get detailed information about your hardware and software configuration. You can find more details about this command at https://developer.toradex.com/software/development-resources/tdx-info/.

Initial Observations

Based on your error message, it appears the container image isn’t being pushed to the local registry correctly before the pull operation attempts to retrieve it. This is a common issue that can be related to:

  • Docker configuration between your development environment and the target device
  • Local registry setup issues
  • Architecture compatibility problems

Docker Desktop Considerations

I notice you mentioned Docker Desktop. The Torizon documentation recommends using Docker Engine instead of Docker Desktop for development (1). Docker Desktop is licensed under the Docker Subscription Service Agreement, while Docker Engine is licensed under the Apache License, Version 2.0 (1). The Torizon IDE team does not support or test the extension flow with Docker Desktop, and you may encounter issues that are not known (2).

Next Steps

Once you provide your system information via tdx-info, I can give you specific guidance for your hardware platform and OS version. The solution will likely involve checking your Docker configuration and ensuring proper communication between your development environment and the target device’s local registry.

Best regards, Toradex AI Assistant.


Did you like my answer? :+1: :-1:

If I answered your question, please tick the “Solution” checkbox.

For follow-up questions to me, please use @ToradexAI in your post.

Bootloader: U-Boot

Kernel version: 6.6.101-7.4.0-g557af2783d91 #1-Torizon SMP PREEMPT Thu Sep 25 07:49:28 UTC 2025
Kernel command line: root=LABEL=otaroot rootfstype=ext4 quiet logo.nologo vt.global_cursor_default=0 plymouth.ignore-serial-consoles splash fbcon=map:3 ostree=/ostree/boot.1/torizon/f3a4c0d28279618ce1a13ed03824ad923d19df32c7a6a7097a35bdde21420134/0
Distro name: NAME=“Torizon OS”
Distro version: VERSION_ID=7.4.0-build.28
Distro variant: VARIANT=“Docker”
Hostname: verdin-imx8mp-15677988

Hardware info

HW model: Toradex Verdin iMX8M Plus WB on Dahlia Board
Toradex version: 0058 V1.1B
Serial number: 15677988
Processor arch: aarch64

Thank you for providing your system information! Now I can help you troubleshoot this container build issue on your Verdin iMX8M Plus with Torizon OS 7.4.0 running on the Dahlia Carrier Board.

Root Cause Analysis

The error manifest for localhost:5002/opencvtest-debug:arm64 not found: manifest unknown indicates that the container image wasn’t successfully pushed to the local registry before the pull operation attempted to retrieve it(1). This is a common issue with the container development workflow.

Docker Desktop Issues

As mentioned in my previous response, Docker Desktop can cause compatibility issues with the Torizon development workflow. The Torizon documentation specifically notes that Docker Desktop is not supported or tested with the IDE extension, and you may encounter unknown issues(2).

Recommended Solutions

1. Check Local Registry Status

First, verify that your local registry is properly running on your target device:

curl http://192.168.12.160:5002/v2/_catalog

This should return a JSON response showing available repositories.

2. Verify Docker Configuration

Ensure your Docker host configuration is correctly pointing to your target device:

DOCKER_HOST=192.168.12.160:2375 docker info

3. Manual Build and Push Process

Try building and pushing the container manually to isolate the issue:

  1. Build the container locally
  2. Tag it for your registry
  3. Push to the local registry
  4. Then attempt the pull operation

4. Check Network Connectivity

Verify that your Windows development machine can properly communicate with the Verdin module on both ports 2375 (Docker daemon) and 5002 (local registry).

Alternative Approach

Consider switching from Docker Desktop to Docker Engine for better compatibility with the Torizon development workflow. This is the recommended approach for Torizon development and may resolve the issues you’re experiencing.

Next Steps

  1. Run the registry check command above
  2. Verify your network connectivity to both Docker ports
  3. Try the manual build/push process
  4. Consider migrating from Docker Desktop to Docker Engine

The combination of your Verdin iMX8M Plus V1.1B and Torizon OS 7.4.0 is a well-supported configuration(3), so this issue is likely related to the development environment setup rather than the target hardware.

Best regards, Toradex AI Assistant.


Did you like my answer? :+1: :-1:

If I answered your question, please tick the “Solution” checkbox.

For follow-up questions to me, please use @ToradexAI in your post.

1, 2, and 4 were successful. If it is not recommended to use Docker Desktop, then please help me to install just the docker engine. Remember, I’m new to Linux

Hey @611NW ,

If you go to our install page, there is a drop down for installing manually. You’ll find instructions for this here.