Can I use virtual machine to install jetpack on Tx2?

Hello,I’m Andy.I come from Taiwan. I want to ask a question. Can I use virtual machine to install jetpack on Tx2?

VMs are not supported. The main issue is that the Jetson’s USB repeatedly disconnects and reconnects during a flash, and VMs tend to lose the USB during that process. This does not mean you couldn’t make it work, and some people do get it to work. Most likely though you’ll need to adjust to make sure that USB port is completely owned by the VM. If you don’t mind the trouble, then you could try, but an actual Ubuntu install is more likely to work.

Thanks for your help.I want to ask another question. Can I use Vm to install Jetpack 4.3 on Tx2?Because I use Vm to install jetpack 3.3 on Tx2,it is not sucessful.

A VM is never supported. A VM could work, but you’d have to experiment and make sure the USB is always passed through correctly. I’d say you’d just have to try it, and dig into your VM’s docs to make sure the USB device is 100% owned by the VM no matter how much it disconnects or reconnects. This in turn would depend on which VM you use.

I used SDK manager to install TX2.But I input my Id which is on my ethernet’s computer.User name is nvidia .Passward is nvidia .

I am not certain what the question is from this, but some details which might help you or others follow below…

JetPack/SDK Manager has basically two modes. The first is to flash, and for this networking is not used between host and Jetson (although SDKM needs access to internet). The flash is in recovery mode, and the Jetson becomes a custom USB device. No network runs between PC and Jetson. During the flash USB connects and disconnects several times. Once the flash is complete, then the Jetson reboots itself, and first login is used to create an administrator login and password (presumably you created the user login “nvidia”, password “nvidia”…this used to be created by default in older releases, but to have this in current releases you would have actually had to set this up manually on first boot).

The second mode of the SDKM is to use networking to install extra/optional packages, e.g., CUDA. For this step to succeed the host computer must be able to reach the IP address of the Jetson, and then log in as the given user with the password you provide. Any failure of ssh, and any failure of the account credentials, would imply failure to install extra software.

If packages fail, then you can try to manually ping the Jetson IP address, and if this works, then you can attempt to manually login via ssh to your account (in this example “ssh nvidia@192.168.50.219”). If this works, then SDKM should also work. However, there is one other requirement sometimes overlooked.

The host computer itself must have access to the internet to download the various packages. The Jetson will need access to the internet if package updates are to occur. Any package conflict which cannot be resolved via internet access (either via host PC access or Jetson access) would block progress. So make sure that both host PC and Jetson can also access the internet. In cases where you get stuck at some stage after the internet is needed, then you should question if internet access is being blocked (this includes firewall or router issues). If you can at least ssh to the Jetson at this stage (and during optional package install there is no reason to not “ssh nvidia@192.168.50.219” as a test), then very likely the internet is an issue.