Hi,
Now I start to run FoundationPose installation.
When I activate the Isaac ROS environment and install the prebuilt Debian package:
isaac-ros activate
sudo apt-get update
sudo apt-get install -y ros-jazzy-isaac-ros-foundationpose
In final I got message:
Selecting previously unselected package ros-jazzy-isaac-ros-foundationpose.
Preparing to unpack ...
/526-ros-jazzy-isaac-ros-foundationpose_4.4.0-0noble.20260428063034213_amd64.deb ...
Unpacking ros-jazzy-isaac-ros-foundationpose (4.4.0-0noble.20260428063034213) ...
Selecting previously unselected package i965-va-driver:amd64.
Preparing to unpack ...
/527-i965-va-driver_2.4.1+dfsg1-1ubuntu0.1_amd64.deb ...
Unpacking i965-va-driver:amd64 (2.4.1+dfsg1-1ubuntu0.1) ...
Selecting previously unselected package va-driver-all:amd64.
Preparing to unpack ...
/528-va-driver-all_2.20.0-2ubuntu0.1_amd64.deb ...
Unpacking va-driver-all:amd64 (2.20.0-2ubuntu0.1) ...
Selecting previously unselected package vdpau-driver-all:amd64.
Preparing to unpack ...
/529-vdpau-driver-all_1.5-2build1_amd64.deb ...
Unpacking vdpau-driver-all:amd64 (1.5-2build1) ...
Selecting previously unselected package proj-bin.
Preparing to unpack ...
/530-proj-bin_9.4.0-1build2_amd64.deb ...
Unpacking proj-bin (9.4.0-1build2) ...
Errors were encountered while processing:
/tmp/apt-dpkg-install-dGuNUh/078-libglvnd0_1.7.0-1build1_amd64.deb
/tmp/apt-dpkg-install-dGuNUh/097-libglx0_1.7.0-1build1_amd64.deb
/tmp/apt-dpkg-install-dGuNUh/098-libgl1_1.7.0-1build1_amd64.deb
/tmp/apt-dpkg-install-dGuNUh/116-libopengl0_1.7.0-1build1_amd64.deb
/tmp/apt-dpkg-install-dGuNUh/118-libegl1_1.7.0-1build1_amd64.deb
/tmp/apt-dpkg-install-dGuNUh/288-libgles1_1.7.0-1build1_amd64.deb
/tmp/apt-dpkg-install-dGuNUh/289-libgles2_1.7.0-1build1_amd64.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)
Could someone help me solve this issue? Thanks in advance.
My temporarily workaround is directly access docker image nvcr.io/nvidia/isaac/ros instead of isaac-ros activate, without granting GPU access. After installng ros-jazzy-isaac-ros-foundationpose I save container to new image, and use the new image to create container with granting GPU access, and use container now to test FoundationPose.
I’ll keep testing this workaround next week, if it seems OK I’ll close this issue.
(Also I found that official guide transforming tensorrt files uses fp32. After add fp16 flag I can run ROS bag example with my 8GB GPU.)
Hello @s5078345,
Welcome to the Isaac ROS forum and thanks for the post!
Are you trying to install the package on X86_64 platform?
I’ve tested the same package you’re trying to install on my environment and it appeared to be installable, so it might not be the package Debian issue.
The log you shared showing that most of the failure are related to OpenGL / EGL / GLVND.
These are graphics dispatch/runtime libraries, not FoundationPose-specific code.
Could you have a quick check inside your container to see if you can access the gpu correctly with the command nvidia-smi.
Then check what’s the current state of the GL libs inside your container.
dpkg -l 'libnvidia-gl-*' 'libglvnd0' 'libgl1' 2>/dev/null
You could provide the detail of below to see what’s the toolkit injecting.
docker run --rm --gpus all ubuntu:24.04 bash -lc '
ls -la /usr/lib/x86_64-linux-gnu/libGL.so* /usr/lib/x86_64-linux-gnu/libEGL.so* 2>/dev/null
dpkg -S /usr/lib/x86_64-linux-gnu/libGL.so.1 2>&1 | head -2
grep -E “libGL|libnvidia” /proc/mounts | head -5
'
Thanks for the reply.
I’m running on x86_64 platform (ubuntu 24.04) with a single NVIDIA GeForce RTX 4060 Ti.
I can access the gpu correctly inside container:
(base) sam@sam-Ubuntu:~$ isaac-ros activate
/home/sam/workspaces/isaac_ros-dev
Logged in to nvcr.io/nvidia/isaac/ros. Using this for cache.
['isaac_ros']
Dockerfile created: image_key = isaac_ros
isaac_ros_28556f8bc78a98822bd08b2d7c6fcf9b-amd64: Pulling from nvidia/isaac/ros
Digest: sha256:c6f5954e318babdcfd152584e03ef68392f309579153f2cd25f788ef7a7046f0
Status: Image is up to date for nvcr.io/nvidia/isaac/ros:isaac_ros_28556f8bc78a98822bd08b2d7c6fcf9b-amd64
nvcr.io/nvidia/isaac/ros:isaac_ros_28556f8bc78a98822bd08b2d7c6fcf9b-amd64
Using image: nvcr.io/nvidia/isaac/ros:isaac_ros_28556f8bc78a98822bd08b2d7c6fcf9b-amd64
Using additional Docker run arguments from ['/etc/isaac-ros-cli/.isaac_ros_dev-dockerargs']
['-v /home/sam/.ssh:/home/admin/.ssh:ro', '-v /home/sam/.aws:/home/admin/.aws:ro', '-v /home/sam/.cache:/home/admin/.cache', '-v /home/sam/.config:/home/admin/.config', '-v /home/sam/.gitconfig:/home/admin/.gitconfig', '-v /home/sam/.bash_history:/home/admin/.bash_history']
Running isaac_ros_dev_container
Creating non-root container 'admin' for host user uid=1000:gid=1000
cp: cannot create regular file '/home/admin/.bashrc': Read-only file system
cp: cannot create regular file '/home/admin/.profile': Read-only file system
chown: changing ownership of '/home/admin/.bashrc': Read-only file system
chown: changing ownership of '/home/admin/.profile': Read-only file system
To run a command as administrator (user "root"), use "sudo <command>".
See "man sudo_root" for details.
admin@sam-Ubuntu:/workspaces/isaac_ros-dev$ nvidia-smi
Tue May 19 08:43:36 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.126.09 Driver Version: 580.126.09 CUDA Version: 13.0 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Persistence-M | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 4060 Ti Off | 00000000:01:00.0 On | N/A |
| 0% 38C P5 17W / 160W | 858MiB / 8188MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
+-----------------------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=========================================================================================|
| No running processes found |
+-----------------------------------------------------------------------------------------+
admin@sam-Ubuntu:/workspaces/isaac_ros-dev$ dpkg -l 'libnvidia-gl-*' 'libglvnd0' 'libgl1' 2>/dev/null
admin@sam-Ubuntu:/workspaces/isaac_ros-dev$
And below is the log following your instructions:
(base) sam@sam-Ubuntu:~$ docker run --rm --gpus all ubuntu:24.04 bash -lc '
ls -la /usr/lib/x86_64-linux-gnu/libGL.so* /usr/lib/x86_64-linux-gnu/libEGL.so* 2>/dev/null
dpkg -S /usr/lib/x86_64-linux-gnu/libGL.so.1 2>&1 | head -2
grep -E “libGL|libnvidia” /proc/mounts | head -5
'
lrwxrwxrwx 1 root root 11 May 19 00:48 /usr/lib/x86_64-linux-gnu/libEGL.so -> libEGL.so.1
lrwxrwxrwx 1 root root 15 May 19 00:48 /usr/lib/x86_64-linux-gnu/libEGL.so.1 -> libEGL.so.1.1.0
-rw-r--r-- 1 root root 72312 Apr 8 2024 /usr/lib/x86_64-linux-gnu/libEGL.so.1.1.0
lrwxrwxrwx 1 root root 10 May 19 00:48 /usr/lib/x86_64-linux-gnu/libGL.so -> libGL.so.1
lrwxrwxrwx 1 root root 14 May 19 00:48 /usr/lib/x86_64-linux-gnu/libGL.so.1 -> libGL.so.1.7.0
-rw-r--r-- 1 root root 547136 Apr 8 2024 /usr/lib/x86_64-linux-gnu/libGL.so.1.7.0
dpkg-query: no path found matching pattern /usr/lib/x86_64-linux-gnu/libGL.so.1
bash: line 4: $'libnvidia\342\200\235': command not found
It seems like the problem is when you launch the container:
- It takes the host’s NVIDIA GL libraries
- Bind-mounts them into the container at the standard paths
(e.g., /usr/lib/x86_64-linux-gnu/libGL.so.1.7.0)
- Does NOT register them in the container’s dpkg database
That’s why the file is there but dpkg has no record of it and show dpkg-query: no path found matching pattern.
Could you share the output of this command form inside the Isaac ROS container?
dpkg -l libopencv-dev libopencv-core4t64 libgl1 libegl1 libglvnd0 libnvidia-gl-580 libnvidia-compute-580 2>/dev/null | grep -E “^ii”
Your save-and-relaunch workaround seems to be a valid fix. But if you’d rather stay in your existing isaac-ros activate session while we work through this, you could use ‘dpkg-divert’ to redirect the conflicting package files away from the toolkit-injected paths before installing foundationpose:
ls /usr/lib/x86_64-linux-gnu/lib{GL,EGL,GLES,GLX,GLdispatch,OpenGL}.so2>/dev/null \
| xargs -I{} sudo dpkg-divert --add --no-rename --divert “{}.distrib” “{}” \
&& sudo apt install -y ros-jazzy-isaac-ros-foundationpose
Let me know if it works for you!
Could you share the output of this command form inside the Isaac ROS container?
dpkg -l libopencv-dev libopencv-core4t64 libgl1 libegl1 libglvnd0 libnvidia-gl-580 libnvidia-compute-580 2>/dev/null | grep -E “^ii”
admin@sam-Ubuntu:/workspaces/isaac_ros-dev$ dpkg -l libopencv-dev libopencv-core4t64 libgl1 libegl1 libglvnd0 libnvidia-gl-580 libnvidia-compute-580 2>/dev/null | grep -E “^ii”
admin@sam-Ubuntu:/workspaces/isaac_ros-dev$
Nothing is listed under command.
Your save-and-relaunch workaround seems to be a valid fix. But if you’d rather stay in your existing isaac-ros activate session while we work through this, you could use ‘dpkg-divert’ to redirect the conflicting package files away from the toolkit-injected paths before installing foundationpose:
ls /usr/lib/x86_64-linux-gnu/lib{GL,EGL,GLES,GLX,GLdispatch,OpenGL}.so2>/dev/null \
| xargs -I{} sudo dpkg-divert --add --no-rename --divert “{}.distrib” “{}” \
&& sudo apt install -y ros-jazzy-isaac-ros-foundationpose
After executing these commands first, I can successfully execute sudo apt install -y ros-jazzy-isaac-ros-foundationpose.
Thanks for your help! I think we can close this topic.