Hello,
I am working on creating a custom image for RevPi Core 3. I am using imagebakery repo on Github.
From my understanding, the image can only be created by running it on a RevPi model. However, this would not work in the scenario of mass deploying a fleet of devices.
What is your recommendation to have a similar setup like building an image with Buildroot/Yocto and flash those on eMMC. That would help us the need to log in to each device, download image bakery repo and run the customize_image.sh command.
Thanks,
Alex
Proper way to creating custom image
Re: Proper way to creating custom image
The customize_image.sh script in the imagebakery repo is intended to be applied to a Raspbian image. It is run once and the resulting image can then be mass-flashed to a multitude of RevPis.
You do not have to run customize_image.sh on a RevPi. It is perfectly fine to run it on a faster x86 machine. You just need to make sure that you run "apt-get install qemu-user-static binfmt-support" as root beforehand (assuming the x86 machine runs Debian).
Of course it is alternatively possible to build a root filesystem with yocto, buildroot, ptxdist or any other common framework, but unfortunately we do not have a "cut and dried" recipe for one of these at hand. We chose to ship the products with a customized Raspbian because that's what people are familiar with if they've used the Raspberry Pi before. That way, their learning curve is as gentle as possible. There are however numerous yocto examples for the Raspberry Pi out there as a quick Google search shows. One thing to consider is that you'll need the devicetree files as well as kernel support for the RevPi-specific chips, such as the RTC and the Ethernet controllers on the PiBridge. The easiest way to achieve that is to just extract our "raspberrypi-kernel" deb-package to the root filesystem, but it's also possible to compile and install the kernel and piControl yourself in the same way our kernelbakery does.
Does this answer your question? If not please try to rephrase it because I'm not sure I've understood correctly what your needs are. Thanks.
You do not have to run customize_image.sh on a RevPi. It is perfectly fine to run it on a faster x86 machine. You just need to make sure that you run "apt-get install qemu-user-static binfmt-support" as root beforehand (assuming the x86 machine runs Debian).
Of course it is alternatively possible to build a root filesystem with yocto, buildroot, ptxdist or any other common framework, but unfortunately we do not have a "cut and dried" recipe for one of these at hand. We chose to ship the products with a customized Raspbian because that's what people are familiar with if they've used the Raspberry Pi before. That way, their learning curve is as gentle as possible. There are however numerous yocto examples for the Raspberry Pi out there as a quick Google search shows. One thing to consider is that you'll need the devicetree files as well as kernel support for the RevPi-specific chips, such as the RTC and the Ethernet controllers on the PiBridge. The easiest way to achieve that is to just extract our "raspberrypi-kernel" deb-package to the root filesystem, but it's also possible to compile and install the kernel and piControl yourself in the same way our kernelbakery does.
Does this answer your question? If not please try to rephrase it because I'm not sure I've understood correctly what your needs are. Thanks.
