When can I use modules for ARM 64 ?

[tkanoh]

We got following error message when start dokers of IoT Edge on ARM64 device.
So, When can I use modules for ARM 64.

iotedgctl start

.............

INFO: Please note depending on network conditions and registry server availability this may take a few minutes.
ERROR: Error during pull for image microsoft/azureiotedge-agent:1.0-preview
404 Client Error: Not Found ("no such image: microsoft/azureiotedge-agent:1.0-preview: No such image: microsoft/azureiotedge-agent:1.0-preview")
ERROR: Exiting with errors. Return code: 1
ERROR: Exiting with errors. Return code: 1

root@obsgold:~# docker pull microsoft/azureiotedge-agent:1.0-preview
1.0-preview: Pulling from microsoft/azureiotedge-agent
no matching manifest for linux/arm64 in the manifest list entries

[darobs]

Hello @tkanoh

We don't currently have ARM64 images for the Azure-IoT-Edge, however, this is a feature on the roadmap. From our experience, the ARM32 images will work on ARM64 architectures.

It looks like Docker is not automatically choosing the ARM32 images, which is annoying. I'm going to write a bug on our backlog to see if we can get the manifest tweaked to choose the ARM32 images for ARM64.

In the meantime, you can work around this by specifying the arm32 images by their tag, "1.0.0-preview022-linux-arm32v7".

edgeAgent: microsoft/azureiotedge-agent:1.0.0-preview022-linux-arm32v7
edgeHub: microsoft/azureiotedge-hub:1.0.0-preview022-linux-arm32v7
tempSensor: microsoft/azureiotedge-simulated-temperature-sensor:1.0.0-preview022-linux-arm32v7

iotedgectl has a --image option in the setup, and the Azure Portal has the ability to set the edgeHub and tempSensor images.

[tkanoh]

Thanks for reply.

I know that ARM32's objects will work on ARM64, but as you pointed out,
Docker does not have the ability to automatically chose the ARM32 image for ARM64.

However, in the meantime I think that we can deal with the --image option.
Thank you for the information.

[krunallokhande]

Hi,

Are the IoT Edge GA arm32 images compatible with arm64 images? We are facing issues installing it on Nvidia Jetson TX2 hardware with arm64 architecture.