Cloning TX1 image on different host machine

[bobtheblob]

Hi all, I’m trying to clone a saved TX1 eMMC image. I’ve been able to successfully clone using the same host computer that the copy was made on. Following the instructions here:

https://devtalk.nvidia.com/default/topic/898999/tx1-r23-1-new-flash-structure-how-to-clone-/

However I am trying to parallelise the TX1 flashing process, which requires me to take the copied image and clone it from multiple host computers (not just the host the clone image was copied from). When I try to copy the ‘system.img’ file onto a different host computer (which also has the SAME VERSION of Jetpack installed), the command:

sudo flash.sh -r -S 14580MiB jetson-tx1 mmcblk0p1

appears to work fine - but when it’s completed, the TX1 does not boot.

I have verified that the same ‘system.img’ when flashed from the original host computer on to the TX1, everything works fine!

Has anyone else encountered this or is trying to do something similar? Is there a file in addition to ‘system.img’ which needs to be copied in order for the cloning to be successful?

Thanks all!

[linuxdev]

Beware that when restoring a clone to a partition that the other partitions have to have a layout exactly the same as the Jetson the clone came from. So if the size of the root partition on the unit being restored to differs, then it won’t work. If the release version differs, e.g., R24.1 clone from and R24.2 restore to, then it may also fail. On top of all this, also beware that no two recovery mode operations can occur in series without restarting recovery mode between operations…e.g., if you clone a Jetson, then the Jetson must be put back into recovery mode before a flash or restore operation can be successful (the Jetson would already be in recovery mode, but the state must be reset for this other action).