Jetson TX2/ XAVIER production OS/SW custom images


I would like to know what is the process to build a custom image with custom SW (i.e. our application that will run on a Jetson) and then create an image that includes the OS with our settings, etc, to then use it for our product

then how can we copy these images into many Jetsons modules at once ?

is there any process documented?

+1, I would love to know how to do this as well (for Xavier)

The easiest way is to build an actual system with everything set the way you want, and clone this (so for example you can update packages, add some, remove some, and the clone will have those updates or edits). The clone can be installed instead of the sample rootfs. If there are device tree changes, then this is a separate edit. In some cases you may have to tweak if for example you put something in which is dependent upon a network MAC address since the other hardware will have a different MAC. See:

Depending on release some details will differ.

Great, while I appreciate the help, it doesn’t answer the whole question, as the other part of the question is how once an image is created, this can be deployed to the Jetson as serial production (i.e. 20 at the time)

does Nvidia has a replicator or a burner we can use to deliver the custom OS, software, etc


Sorry, that’s one I can’t help with. I know in the past there were documentation sections on flashing multiple devices simultaneously, but for the R32.x+ I’m not sure.

hello henry.acevedo,

you should make a golden build locally by clone the image.
suggest you also refer to Topic 1038136 for the script to flash several tegra device simultaneously,