new TX2 dev quit : "bzip2: Data integrity error" at initial installation


running the nvidia installer on new TX2 dev kit, I’m getting an error on BSP extraction

[code]Last login: Wed Sep 11 13:30:15 UTC 2019 on tty1
Welcome to Ubuntu 16.04 LTS (GNU/Linux 4.4.38-tegra aarch64)

717 packages can be updated.
441 updates are security updates.


These are instructions for installing the NVIDIA Linux driver
binary release which is located at: “${HOME}/NVIDIA-INSTALLER”

Step 1)
Change directories into the NVIDIA installation directory:

Step 2)
Run the installer script to extract and install the NVIDIA
Linux driver binary release:
sudo ./

Step 3)
Reboot the system to have the Ubuntu Desktop UI come up.

username: nvidia
password: nvidia

Visit to
download the latest software release and product documentation

nvidia@tegra-ubuntu:~ cd Hae[KNVIDIA-INSTALLER/ nvidia@tegra-ubuntu:~/NVIDIA-INSTALLER ls Linux_for_Tegra Tegra186_Linux_R28.1.0_aarch64.tbz2
nvidia@tegra-ubuntu:~/NVIDIA-INSTALLER$ suoe[Kdo ./
[sudo] password for nvidia:
Extracting the BSP…

bzip2: Data integrity error when decompressing.
Input file = (stdin), output file = (stdout)

It is possible that the compressed file(s) have become corrupted.
You can use the -tvv option to test integrity of such files.

You can use the `bzip2recover’ program to attempt to recover
data from undamaged sections of corrupted files.

tar: Unexpected EOF in archive
tar: Unexpected EOF in archive
tar: Error is not recoverable: exiting now
ERROR extracting Tegra186_Linux_R28.1.0_aarch64.tbz2
nvidia@tegra-ubuntu:~/NVIDIA-INSTALLER$ sudo halt
[sudo] password for nvidia:
[ 67.858627] -->eqos_shutdown
[ 67.863016] Handle the shutdown
[ 67.867663] >–eqos_shutdown
[ 67.967695] reboot: System halted

I’m guessing that you’re using a very old release. You might try a newer one. FYI, if your host PC ran out of disk space during a flash, then the image would be truncated, but the flash would continue without telling you.

Which L4T or JetPack release are you using? From the host you ran the flash, and from its “Linux_for_Tegra/” directory, what do you see from:

cd /where/ever/it/is/Linux_for_Tegra/
df -H -T .


I’ve bougth the tx2 dev kit from nvidia website last week.

It was already flashed.
On the TX2 I’ve this directory :

nvidia@tegra-ubuntu:~/NVIDIA-INSTALLER$ ls -lh
total 94M
-rwxr-xr-x 1 root   root   2.8K Jul 28  2017
-rw-r--r-- 1 root   root   5.8K Jul 28  2017 LICENSE
drwxr-xr-x 3 nvidia nvidia 4.0K Sep 11 13:28 Linux_for_Tegra
-rw-r--r-- 1 root   root    492 Jul 28  2017 README.txt
-rw-r--r-- 1 root   root    94M Jul 28  2017 Tegra186_Linux_R28.1.0_aarch64.tbz2

I’ll reflash the board, but this post is to inform nvidia team of the issue


Interesting that the default shipped unit would have this issue. Back when R28.1 was released the Ubuntu 18.04 bug with bzip2 did not yet exist, so it is unlikely the bzip2/bunzip2 is related to the data integrity failure.

This is more or less trivia you may find interesting about the “” script. I wouldn’t bother with the information except that you ran into this error…

During a normal flash the sample rootfs is pure Ubuntu without NVIDIA drivers. That Ubuntu content is kept separate from the NVIDIA content for licensing reasons. The base installation, if the NVIDIA content has not been installed, will lack direct hardware access, e.g., CUDA won’t work. In a normal flash there is a script, “sudo ./”, and this script writes the NVIDIA content into the Ubuntu sample rootfs prior to flash. After flash the content is in place. If you have a factory flashed unit, then the “” does exactly the same thing by copying the NVIDIA drivers into the running system (and after this CUDA can work).

By flashing you won’t have to worry about that content failing to install. R28.1 is also rather old, and so you are advised to upgrade anyway. The most recent release of SDK Manager does seem to have an awareness of a bunzip2 bug in one of the Ubuntu releases (host side), and so you shouldn’t run into any truncation or integrity issues using the most recent release.

On any of the running Jetsons you can use this command to test the integrity of the NVIDIA-specific infrastructure content:

sha1sum -c /etc/nv_tegra_release