[L4T 32.3.1] Does not create GPT partition during mass flashing

Hello,

We are running a mass flashing station of five production modules. We have our mass img created from ./nvmassflashge.sh available. The issue is that during mass flashing (i.e. ./nvmflash.sh), sometimes one of the boards will be stuck creating the GPT partition. It would be stuck on this line, even after 10min of inactivity:

Writing partition GPT with gpt.bin

I just wanted to know what could possibly cause this issue during flashing, and if there’s anything we can do to mitigate other errors like these.

We have an Intel NUC connected to an external USB hub connected to five Gumstix FastFlash boards (https://store.gumstix.com/manufacturer/jetson-nano-flasher.html), and we attach the production modules to the FastFlash boards.

If USB had an issue, then this could cause symptoms not looking like USB (though I’d think dmesg on the host would mention a USB error…the flash software probably would not mention USB in some cases). I have heard of some Intel NUCs having USB problems, so you might test with a different host PC. Also monitor “dmesg --follow” during flash, and when notice something slowing down, see if dmesg mentions any kind of error or issue.

20200915-143459_2595_flash_3-4.2.log (3.4 KB) 20200915-143900_3022_flash_3-2.log (8.8 KB)

The logs indicate operations up until the flash of APP (rootfs) succeed, and then that partition never proceeds. There could be an actual hardware failure, and as long as you are not using a VM, it is difficult to say if the failure is at the Jetson, host PC, or cable.

Does the error follow a single Jetson? Or a single cable? Or a single host PC?