Support x86_64. Video analyze services (Not IoT)

Hello,

I am looking for the feasibility of OpenBalena to deploy and manage my services(Docker-based, Video analyze, Not IoT) which are running on Cuda enabled GPU machines on multiple sites. So, I just have gone through brief OpenBalena documents. However, it seems x86_64 is limited to use OpenBalena.

Host Machine: x86_64(Intel i7 Core), Nvidia GPUs, Ubuntu 18.04

Is there any workaround for this case? I would appreciate it if you could help or add any comment to go forward and try it out.

Thanks, LimPro

Hi LimPro,

Welcome to the forums! Your use case sounds really interesting, so if you don’t mind I’d like to ask for some more details before we try and point you in the right direction!

Can you tell me more about your machine set up, and what it is you would like to do? Do I understand that you would like to use openBalena rather than balenaCloud, and you would like to run balenaOS on your GPU-enabled devices that are doing the video analysing? Is there a reason you don’t want to use the balenaCloud? What are the specs of your GPU machines? And what would be the spec of the machine you’d like to run openBalena on?

If you could let me know more about what you would like to be able to do, we’d love to help as much as we can!

Phil

Hi Phil,

Thank for your reply.

Basically the machines have 8~12 CPU core(Intel i7), 32~64GB memory and 2~3 NVIDIA GPU for video analysing and Ubuntu18.04 is currently running on the machines.
I would like to use most of openBalena key components and features including balenaCloud, dashboard, balenaOS(on GPU enabled machines if it is possible)

LimPro

Hi again - so just to be clear, before I go and ask the openBalena experts:

  • You will have device running openBalena - what would be the specs of that machine?
  • You have Ubuntu machines with multiple NIVIDIA GPUs, which you would want to install balenaOS on the hosts, containerise the video analsis software, and deploy those to the openBalena machine?

Thanks
Phil

Hi Phil,

I added machine info as below:

  • You will have device running openBalena - what would be the specs of that machine?
    => I would be x86_64(i7), 32GB memory,

  • You have Ubuntu machines with multiple NIVIDIA GPUs, which you would want to install balenaOS on the hosts, containerise the video analsis software, and deploy those to the openBalena machine?
    => Yes. 8~12 CPU core(Intel i7), 32~64GB memory and 2~3 NVIDIA GPU for video analysing and Ubuntu18.04 is currently running on

Thanks, LimPro

Hey @limpro,

There seems to be a bit of confusion with the different products. Let me try to offer a quick explanation and see if we can figure out how to help you.

openBalena is the open source version of balenaCloud. They share the same core technology, but some features are only available on balenaCloud (like the web dashboard). You can find more information about the differences here.

balenaOS is our own operating system optimized to running docker and allows devices to be connected to openBalena or balenaCloud. We currently support some x86 devices (Intel NUCs for example) but given the specs you shared, custom hardware support might be required.

I’ll reach out to the OS team to see what would be required to support your HW, but the information above should help you understand if you need balenaCloud or openBalena.

Cheers,
Nico

Hi Nico,

Thanks for your pointing out. I am interested in the dashboard feature(seems only supported from balenaCloud) and other deploying feature. Also, I would like to try both out balenaCloud and OpenBalena if balenaOS is available on x86_64.
I am looking forward to having update from OS team.

Thanks, LimPro

Hey there!

You are more than welcome to sign-up for a balenaCloud trial. Notice that you can sign-up for an enterprise plan that would get you an on-prem managed balena installation that you can self host: https://www.balena.io/pricing.

Let us know if you have any questions!

Hi Jviotti,

Thanks for your comment.

I found “Generic x86_64” on the device type selection in https://dashboard.balena-cloud.com/apps
However, I could found detailed information on this “Generic x86_64”

Can you help me figure it out, please?

Thanks, LimPro

Hi, we support 2 types of x86 64 bits machines: 1 is for the Intel NUC variants and 1 is similar but it supports a wider variant of drivers and it is called Generic x86_64. Given your hardware, I would advise you use the second machine type. So please give it a try and let us know how it goes

1 Like

Hi Floion,

Thanks for your comment.
could you help correct version of the generic x86_64 image version to download(by balenaCLI), please?
As I attached my screenshot, I failed to download balenaOS of generic x86_64 image. please kindly collect me if I was wrong.

Thanks, LimPro

Hi
I just tried the command that you were running and it seems to work just fine on my end

 balena os download genericx86-64-ext -o out.img --version latest

Can you try copy pasting the above and trying again? There must be some small typo on your end

Hi Anujdeshpande,

I copied your command and ran it but still failed.

That’s strange - can you please share the output of balena --version and uname -a

Here we go!

FYI, The corresponding OpenBalena is running at AWS instance.

Oh! In that case, OpenBalena doesn’t support os downloads. You will have to download the image from balena.io/os

1 Like

Unfortunately, I cannot find generic86-64-ext at https://www.balena.io/os/

If I run openBalena on my local machines, then I can download OS images?

Hi

Do you have a balenaCloud account? One thing that you can do is that you can configure your CLI to use the credentials for balenaCloud while downloading the image. And then once you have the images you can go back to using OpenBalena

Yes, I created balenaCloud account yesterday. Please let me try it out!

BTW, I figured out that I can download generic-x86-64 os image from below link. Can it be used in openBalena also? such as

balena os configure out.img --app myApp

https://api.balena-cloud.com/download?deviceType=genericx86-64-ext&version=2.48.0+rev5.dev&fileType=.zip