I did the following using Jetson Nano modules (version number 900-13448-0020-000) with a Turing Pi 2.
I used a dual booting laptop configured with Ubuntu 18.04 and Windows 10.
I installed the Nvidia SDK Manager while running Ubuntu 18.04 on the laptop.
My unsuccessful experience using slots 1 and 2 on the Turing Pi 2 was consistent with others. So I flashed each of the Nanos in slot 3, using the following steps.
Using the Turing Pi 2 UI, I configured the USB OTG port to connect with Node 3.
I used a USB type A to type A cable to connect the Turing Pi 2 USB OTG port to a USB port on the laptop.
Using the Turing Pi 2 UI, I powered Node 3 on.
On the laptop, I started the Nvidia SDK Manager and the device was recognized.
I selected JetPack 4.6.2 with which to flash the Nanos. For expedience and a desire to reduce the number of variables at play, I deselected the additional modules, keeping on the base JetPack.
After the SDK manager reported success (this took a while), I powered the newly flashed Nano off and on again, using the Turing Pi UI.
Once the Nano showed up on my network with the hostname 'ubuntu', I ssh'd into it and finished configuration (e.g. hostnamectl -sethostname).
I then powered the slot down again with the Turing Pi UI, moved the Nano to its destination slot, and powered the Nano in its new slot.
Then I started again with the next Nano.