Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other
Target Operating System
Linux
QNX
other
Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
DRIVE AGX Orin Developer Kit (940-63710-0010-100)
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
other
SDK Manager Version
1.9.3.10904
other
Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
other
Hello,
when I’m trying to run the samples from the DriveWorks SDK Reference (e.g. “hello_world_sample”) on the host, I always got the following error …
seybothd@fnl8014u:~$ /usr/local/driveworks/bin/sample_hello_world
/usr/local/driveworks/bin/sample_hello_world: error while loading shared libraries: libEGL.so: cannot open shared object file: No such file or directory
I want do go through this reference, and later on accessing sensors and using cameras.
Btw. why are these examples all located on the host and not on the Drive AGX Orin?
Dear @DominikS ,
We don’t ship samples on target to save filesystem space and suggest you to cross compile if you want to run on target. Please check DW sample cross compilation instructions at DriveWorks SDK Reference: Getting Started Using the NVIDIA SDK Manager
Hello @SivaRamaKrishnaNV,
thanks for the explanation and advice to use the cross compilation. Unfortunately, I get an error during the step “make” and “make install”
/usr/local/driveworks/samples/3rdparty/src/glfw/3.2-screen/src/x11_platform.h:39:10: fatal error: X11/Xcursor/Xcursor.h: Datei oder Verzeichnis nicht gefunden
I attached the log from the host terminal and the build files.
error_during_crosscompile.txt (19.8 KB)
build-aarch64-linux-gnu-DCMAKE_TOOLCHAIN_FILE=.zip|attachment (863.7 KB)
Dear @DominikS,
Is it possible to upgrade to DRIVE OS 6.0.8.1? If possible, you can use DRIVE OS 6.0.8.1 docker container which will resolve issues related to host environment.
Hello @SivaRamaKrishnaNV,
of course, not a problem to update - can I use this Install DRIVE OS 6.0.8.1 for NVIDIA Developer Users | NVIDIA Docs introduction (for NVIDIA Developer Users)?
In that case, the SDK manager approach cannot be used anymore? Because we flashed OS6.0.6 by SDK manager approach.
Meanwhile I also see the discussion about the need of a NVIDIA GPU to run samples - is this only necessary for run samples on the host, or does it also affect NVIDIA DRIVE Orin Target System Linux aarch64 ( DriveWorks SDK Reference: Getting Started Using the NVIDIA SDK Manager)
Yes. Ampere based GPU is required on host to run DW samples. After, cross compilation, DW samples are expected to run on target. Note that, We expect the the docker container is to use for sample development and cross compilation. Running the sample in docker may require additional config settings.
Ok, so I will follow these instructions Install DRIVE OS Linux Docker Containers from NGC | NVIDIA Docs
and doing both steps …#
Hello @SivaRamaKrishnaNV, it seems that an argument is missing when I enter the command …
sudo docker run -it --privileged --net=host -v /dev/bus/usb:/dev/bus/usb -v ${WORKSPACE}:/home/nvidia/
nvcr.io/drive/driveos-sdk/drive-agx-orin-linux-aarch64-sdk-build-x86:latest
I attached this log-file from the terminal
docker_download_not_started.txt (2.2 KB)
Dear @DominikS,
Is this issue resolved?
Hello @SivaRamaKrishnaNV, no unfortunately not. I also checked the docker run | Docker Docs which “Argument” might missing. So far, I don’t see any further hint, where the problem migth be resp. what is missing (I also just copy out the commands from the instructions [ Set Up Docker and NVIDIA GPU Cloud Access | NVIDIA Docs] directly in the comand line of my host PC).
Is there space between /home/nvidia and nvcr.io?
I just make two trials with space and without space inbetween (see the attached picture
Dear @DominikS,
It seems you have not set WORKSPACE variable . Please set using export WORKSPACE=</path/to/workspace> and then run the command.
Ok, I’ll do - is there any suggestion for the workspace?
E.g. should it be the same like the workspace directory in the SDK Manger?
Dear @DominikS,
workspace is a folder in host that gets mapped to a folder in docker(in this is case it maps to /home/nvidia on target). This allows us to share data across docker and host.
@SivaRamaKrishnaNV sry, I have deleted the last post by mistake. Bute here is another screenshot with more progess in the downloading of Docker. However, I wasn’t fully successful yet - it seems that download aborts after a while.
Dear @DominikS,
It looks like network issue. Please check pull the image again and see if other layers are also downloaded.
I repeated the command again since a while and it is still downloading - how many GB in total, seems a huge amount of data!
But it is still running without any issues. I’m just really surprised about the size/amount of data.
After all downloads from Docker are completed, I guess to continue with
Preformatted text./flash.py /dev/ttyACM1 p3710
But further eorrs and no progress.
