Lifestream Failed to acquire rtx::resourcemanager::ResourceManager

I am trying to run the Isaac simulator in docker and lifestream it to my computer. The simulator runs fine until any other process occupies a part of the gpu.

Even if the other process only uses 2Gb out of 50Gb VRAM (Nvidia A6000), I then get this error and a segfault. My guess is that the resource manager tries to lock in the whole GPU. Note that once I have started it, there is no issue to then start other GPU processes afterwards. It would be quite important for us to be able to run other processes on this machine as well. Do you know if there is a way to achieve that?

[2.632s] [ext: omni.rtx.window.settings-0.6.1] startup
[2.636s] [ext: omni.kit.window.stage-2.3.7] startup
[2.639s] [ext: omni.replicator.composer-1.1.3] startup
[2.652s] [ext: omni.kit.browser.core-2.0.12] startup
[2.655s] [ext: omni.rtx.settings.core-0.5.5] startup
[2.660s] [ext: omni.kit.window.extensions-1.1.0] startup
[2.664s] [ext: omni.kit.graph.usd.commands-1.1.0] startup
[2.666s] [ext: omni.kit.browser.folder.core-1.1.13] startup
[2.668s] [ext: omni.isaac.debug_draw-0.1.2] startup
[2.672s] [ext: omni.isaac.ui-0.2.1] startup
[2.673s] [ext: omni.kit.graph.delegate.default-1.0.15] startup
[2.674s] [ext: omni.isaac.motion_planning-0.2.0] startup
[2.679s] [ext: omni.isaac.lula-1.1.0] startup
[2.685s] [ext: omni.isaac.surface_gripper-0.1.2] startup
[2.689s] [ext: omni.kit.graph.editor.core-1.3.3] startup
[2.691s] [ext: omni.isaac.motion_generation-3.6.1] startup
[2.695s] [ext: omni.isaac.manipulators-1.0.1] startup
[2.696s] [ext: omni.isaac.core_nodes-0.13.0] startup
[2.705s] [ext: omni.kit.graph.widget.variables-2.0.2] startup
[2.706s] [ext: omni.kit.graph.delegate.modern-1.6.0] startup
[2.707s] [ext: omni.isaac.franka-0.3.0] startup
[2.708s] [ext: omni.isaac.wheeled_robots-0.5.8] startup
[2.717s] [ext: omni.graph.window.core-1.23.4] startup
[2.722s] [ext: omni.graph.instancing-1.1.4] startup
[2.726s] [ext: omni.graph.tutorials-1.1.2] startup
[2.739s] [ext: omni.graph.window.action-1.3.8] startup
[2.740s] [ext: omni.isaac.isaac_sensor-1.0.2] startup
2022-10-11 14:51:32 [2,738ms] [Warning] [omni.physx.plugin] Deprecated: getSimulationEventStream is deprecated, please use getSimulationEventStreamV2
[2.748s] [ext: omni.kit.selection-0.1.0] startup
[2.749s] [ext: omni.kit.widget.layers-1.5.17] startup
[2.841s] [ext: omni.graph.bundle.action-1.0.0] startup
[2.842s] [ext: omni.kit.profiler.window-1.4.4] startup
[2.846s] [ext:] startup
[2.847s] [ext: omni.isaac.range_sensor-0.4.3] startup
[2.871s] [ext:] startup
[2.873s] [ext: omni.replicator.isaac-1.3.2] startup
[2.959s] [ext: omni.isaac.kit-0.2.1] startup
[2.959s] [ext: omni.graph.window.generic-1.3.8] startup
[2.960s] [ext: omni.isaac.utils-0.1.11] startup
[2.963s] [ext:] startup
[2.964s] [ext:] startup
[2.965s] [ext:] startup
[2.966s] [ext:] startup
[2.967s] [ext:] startup
[2.968s] [ext:] startup
[2.969s] [ext:] startup
[2.972s] [ext: omni.isaac.universal_robots-0.3.0] startup
[2.972s] [ext: omni.isaac.occupancy_map-0.2.4] startup
[2.982s] [ext: omni.kit.window.console-0.2.0] startup
[2.986s] [ext: omni.kit.window.status_bar-0.1.1] startup
[2.989s] [ext:] startup
[2.991s] [ext:] startup
[2.992s] [ext: omni.isaac.dofbot-0.3.0] startup
[2.992s] [ext: omni.kit.streamsdk.plugins-0.0.0] startup
[2.992s] [ext: omni.kit.livestream.core-0.0.0] startup
[2.995s] [ext: omni.kit.window.title-1.1.2] startup
2022-10-11 14:51:32 [2,988ms] [Error] [carb.windowing-glfw.plugin] GLFW initialization failed.
2022-10-11 14:51:32 [2,988ms] [Error] [carb] Failed to startup plugin carb.windowing-glfw.plugin (interfaces: [carb::windowing::IGLContext v1.0],[carb::windowing::IWindowing v1.3]) (impl: carb.windowing-glfw.plugin)
2022-10-11 14:51:32 [2,990ms] [Error] [carb.windowing-glfw.plugin] GLFW initialization failed.
2022-10-11 14:51:32 [2,990ms] [Error] [carb] Failed to startup plugin carb.windowing-glfw.plugin (interfaces: [carb::windowing::IGLContext v1.0],[carb::windowing::IWindowing v1.3]) (impl: carb.windowing-glfw.plugin)
[2.997s] [ext: omni.isaac.urdf-0.4.0] startup
[3.024s] [ext:] startup
2022-10-11 14:51:32 [3,019ms] [Warning] [omni.ext-live.plugin] Unable to detect Omniverse Cache Server. File /root/.nvidia-omniverse/config/Cache/Omniverse.toml is not found. Consider installing it (version >= 280) for better IO performance.
[3.027s] [ext:] startup
[3.028s] [ext: omni.kit.livestream.native-0.1.4] startup
2022-10-11 14:51:32 [3,026ms] [Error] [omni.kit.livestream.plugin] Failed to acquire rtx::resourcemanager::ResourceManager!
[3.033s] [ext: omni.isaac.sim.python-2022.1.1] startup
[3.034s] Simulation App Starting
2022-10-11 14:51:32 [3,097ms] [Error] [] VkResult: ERROR_EXTENSION_NOT_PRESENT
2022-10-11 14:51:32 [3,097ms] [Error] [] vkCreateInstance failed. Vulkan 1.1 is not supported, or your driver requires an update.
2022-10-11 14:51:32 [3,097ms] [Error] [] Vulkan Instance creation failed, or provided an invalid one.
Segmentation fault (core dumped)

Hi. When running a container with the --gpus all flag, the GPU is locked and assigned to the container. You can change this flag to use a specific GPU if you have more than 1 GPU. See User Guide — NVIDIA Cloud Native Technologies documentation

If you have a Multi Instance GPU capable GPU like the A100, you may split it into smaller vGPUs but the A6000 may not support this. See Multi-Instance GPU (MIG) | NVIDIA