Error:CUDA cannot be installed on device.

"Error:CUDA cannot be installed on device.This may be caused by other apt-get command running on device when installing CUDA.Please use apt-get command in a terminal to make sure following packages are installed correctly on device before continue:cuda-toolkit-8-0 libgomp1 libfreeimage-dev libopenmpi-dev openmpi-bin.After these packages are installed on device,press Enter key to continue."Has anybody meets this problems?I am sure that no other apt-get command is running.Otherwise when I flash OS on board My board has no wifi.

Hi SunXiang,

Your issue seems similar as the following thread:
https://devtalk.nvidia.com/default/topic/987525/

Maybe your internet connection failed to pass through a firewall or proxy.
Plesae be sure only to use the wired ethernet during the install.
Thanks

I find an error when installing.

"Return Code: 1
Error: Cannot connect to device with given information. Please check your input and try again.

Detailed info:
Read from socket failed: Connection reset by peer"

Could you please tell me wha’s the reason.I am sure I have installed the ssh service.

Connection reset by peer tends to mean the end point networking received and rejected the connection. A typical scenario for this if you use ssh (and I think in some cases tools use ssh) is that the ssh connection itself is rejected…on a system which was flashed, this is often because keys change each flash unless you preserve them. Can you connect by ssh on command line?

If you cannot connect on command line ssh due to key changes, note that in each receiving end there is a file “~/.ssh/known_hosts”. Each user@host is one very long line (don’t let line wrap fool you). If you delete the line associated with an end point then ssh will ask if you want to allow that connection on the first connect…verifying “yes” puts a new line back in based on current keys. Assuming command line ssh failed before, try again after removing the relevant line. You might need to remove one line from each user@machine involved (host plus Jetson).

Thanks very much.

I get the reason.It is because of JetPack-L4T-2.3.1-linux-x64.run’s priority is too high.When I exec “chmod 711 JetPack-L4T-2.3.1-linux-x64.run” and reload all the file,I success.

遇到相似的一个问题,用jetPack单独更新tx1上的内容时,连接不上tx1:
主机是Ubuntu14.04(虚拟机)和tx1可以ping同,但是主机上需要配置ssh 服务么?不了解ssh使用,伸手·········

Hi Ka-robot-whi,

Welcome to Jetson & Embedded Systems developer forum.
Please create the topic and issue in English manner.
Other developers can understand your issue and to share their experience with you quickly if they ever met similar issue in design stage.
And then other users could be leverage from this issue to get benefits.

For your issue, let’s discuss at below link:
https://devtalk.nvidia.com/default/topic/998232/jetson-tx1/what-is-the-quot-user-name-and-password-quot-means-in-the-jetpack-l4t-doc-/

Thanks