I have tried everything I can find online to try and get SDK manager to let me flash my jetson nano. I have spent 3 weekends installing and un-installing linux OS’s. I’ve tried changing the version to 18.4, i’ve tried to down load 18.4. Why is this so hard? Had my nano for 2 months. I might as well just toss it in the trash as there is no way to get SDK manager to let me down load an image. Its see it! but that’s it. wont let me download the flash image at all. Been trying everything I can try. nothing. Why cant I just go to a download page for the image? why does this have to go though sdk???
Hello troy98133
You can download the Image from this link: https://developer.download.nvidia.com/embedded/L4T/r32_Release_v7.1/JP_4.6.1_b110_SD_Card/Jetson_Nano_2GB/jetson-nano-2gb-jp461-sd-card-image.zip
Follow this instructions to flash the Image in the SD: https://developer.nvidia.com/embedded/learn/get-started-jetson-nano-2gb-devkit#write
See the section Instructions for Linux:
/usr/bin/unzip -p ~/Downloads/jetson_nano_devkit_sd_card.zip | sudo /bin/dd of=/dev/sd<x> bs=1M status=progress
If you need to use the SDK manager I recommend to use a Ubuntu 18.04 Docker image from here: Log in | NVIDIA Developer
Manuel Leiva
Embedded SW Engineer at RidgeRun
Contact us: [email protected]
Developers wiki: https://developer.ridgerun.com
Website: www.ridgerun.com
Just some trivia related to why the flash software…
Jetsons do not have a hardware BIOS. They have the software equivalent, and that means flashing them more or less also flashes the BIOS. The SD card of SD card models (not all models lack eMMC) does not include that early boot content. The early boot content is in QSPI memory on the module itself, and the SD card is unable to influence that. If a Jetson has not been flashed once with the same major release version as that of the SD card content, then it won’t ever find the SD card and won’t boot. For this model you are essentially performing a one time flash of QSPI, after which you won’t need to flash it again when changing SD cards.
Incidentally, JetPack/SDK Manager is just a front end to the actual flash software. Command line doesn’t do everything the GUI does, but for basic flash it is fairly simple and has fewer requirements on the host PC performing the flash.
downloaded file. will see how it goes. Crossing fingers. Working with a 11 year old pc thats windows 10 only with maxed hard drives to research this, and an old antec i5 min pc I found in the trash and got running with salvaged parts to run (now Ubuntu 22.04) with a mostly working 500gb HDD. Thanks for the help! GPT only had links that all 404’ed. It has me leaning more about lunix than I ever thought I would, but as most of what I need for the jetson nano needs to work with lunix, Its just learning curve I have to get through. Got the nano with amazon gift card I got from work. Just a maintenance guy with a disabled husband at home who needs a friend while i’m at work. Out in the woods so to speak. Used Anaconda and GPT to get a blender bot working about a year ago, but to slow on my windows pc. got it to hear me and speak, but when I tried vision, gpt kept having me kill it with version software incompatibility issues. GPT has terrible memory about things we have already done (yes I know thats normal) Thanks again for the help! Will update!
So ya, followed your instructions. I’ve actually tried that image a few times before. Once on a 256gm sd card which I guess is to big. Once on a 64gig micro sd, once on an 8gig micro sd. Nothing, always end up a the Shell prompt. FIgured it might be the version of linux I was using (mint) so tied some version 18 one, and now using Ubuntu 22.04 as I can see that SDK version list that has what I need to do a usb flash. But of course, the Ubuntu 22.04 down load of SDK isn’t any kind of executable I have ever seen before, so having GPT try to help me figure out how to run what I downloaded. will update if I figure anything out, I’m sure I am not the only one having this issue…
ok, trying Ubuntu 20.04 tomorrow and will try SDK manager for 20.04. I mean why not try another OS. All this SDK manager dead end, sd card not recognized, https:// dead links, and missing toolkit ridiculousness is just so fun! Will keep updating so if anyone in the future reads this they might just decide to toss their Jetson nano in the trash. Nvidia abandoned it and apparently deleted everything. Still hoping a solution can be found. I wont be able to afford another Nvidia gpu any time soon. Will be PO if I tossed away $300 for nothing. I still have hope, but starting to feel like a fool lol
Hi troy98133
Please use the Docker image ( Log in | NVIDIA Developer )
docker run -it --privileged -v /dev/bus/usb:/dev/bus/usb/ -v /dev:/dev -v
/media/$USER:/media/nvidia:slave –name jetPack --network host --entrypoint /bin/bash sdkmanager:<tag>
and Use the CLI version of te SDK manager.
sdkmanager --cli install --logintype devzone --product Jetson --version 4.6.4
–targetos Linux --host --target JETSON_NANO_TARGETS --flash all --license
accept --archived-versions
Manuel Leiva
Embedded SW Engineer at RidgeRun
Contact us: [email protected]
Developers wiki: https://developer.ridgerun.com
Website: www.ridgerun.com
You might want to try the Docker image listed above, it is something of an o/s adapter. For an original Jetson Nano, which is running L4T R32.x, then your host PC would be running the flash software using anything based on Ubuntu 18.04. 20.04 will be a problem. Other “flavors” of 18.04 work, e.g., I use the KDE Ubuntu (KUbuntu), but the underlying release would need to be 18.04. Try the above Docker info which can work on other releases of Linux to “pretend” it is 18.04 (or even different distributions), or switch back to an Ubuntu 18.04 derivative.