github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #12173] Unable to load cudart library /usr/lib/x86_64-linux-gnu/libcuda.so.570.169: cuda driver library init failure: 802 #54609

New Issue
Closed
opened 2026-04-29 06:33:19 -05:00 by GiteaMirror · 2 comments
GiteaMirror commented 2026-04-29 06:33:19 -05:00
Owner
Copy Link

Originally created by @connorgoodhue on GitHub (Sep 3, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/12173

What is the issue?

I am new to Ollama, in fact, I am new to all of this. I've kind of been thrown blindly into setting up an "on-prem AI server" at my work. It is a Lambda Hyperplane server running 8 NVIDIA H100s. Anyway, I keep running into this error and I can't get it to go away.

Unable to load cudart library /usr/lib/x86_64-linux-gnu/libcuda.so.570.169: cuda driver library init failure: 802

The "AI server" is running Ubuntu Pro, and Ollama is running inside a Docker CE container (not Snap Docker). I have all the NVIDIA drivers installed at version 570.169 on the host, NVIDIA Container Toolkit is 1.17.8-1. Running CUDA 12.8. The CUDA drivers in the Ollama container are also 12.8. When I run nvidia-smi in the container it matches the versions shown on my host.

I stopped trying to use the official Ollama Docker image, because I thought maybe I could just build off of an NVIDIA Cuda container image for 12.8 and then install Ollama on top of it. But that still didn't work, I get the same 802 error. Here is a more verbose version of the error.

time=2025-09-03T15:13:30.404Z level=INFO source=gpu.go:217 msg="looking for compatible GPUs"
time=2025-09-03T15:13:30.458Z level=DEBUG source=gpu.go:98 msg="searching for GPU discovery libraries for NVIDIA"
time=2025-09-03T15:13:30.458Z level=DEBUG source=gpu.go:503 msg="Searching for GPU library" name=libcuda.so*
time=2025-09-03T15:13:30.458Z level=DEBUG source=gpu.go:527 msg="gpu library search" globs="[/usr/local/lib/ollama/libcuda.so* /usr/lib/x86_64-linux-gnu/libcuda.so* /usr/local/cuda/lib64/libcuda.so* /usr/local/cuda*/targets/*/lib/libcuda.so* /usr/lib/*-linux-gnu/nvidia/current/libcuda.so* /usr/lib/*-linux-gnu/libcuda.so* /usr/lib/wsl/lib/libcuda.so* /usr/lib/wsl/drivers/*/libcuda.so* /opt/cuda/lib*/libcuda.so* /usr/local/cuda/lib*/libcuda.so* /usr/lib*/libcuda.so* /usr/local/lib*/libcuda.so*]"
time=2025-09-03T15:13:30.465Z level=DEBUG source=gpu.go:560 msg="discovered GPU libraries" paths=[/usr/lib/x86_64-linux-gnu/libcuda.so.570.169]
initializing /usr/lib/x86_64-linux-gnu/libcuda.so.570.169
dlsym: cuInit - 0x7fe0dfe48a60
dlsym: cuDriverGetVersion - 0x7fe0dfe48a80
dlsym: cuDeviceGetCount - 0x7fe0dfe48ac0
dlsym: cuDeviceGet - 0x7fe0dfe48aa0
dlsym: cuDeviceGetAttribute - 0x7fe0dfe48ba0
dlsym: cuDeviceGetUuid - 0x7fe0dfe48b00
dlsym: cuDeviceGetName - 0x7fe0dfe48ae0
dlsym: cuCtxCreate_v3 - 0x7fe0dfe48d80
dlsym: cuMemGetInfo_v2 - 0x7fe0dfe69140
dlsym: cuCtxDestroy - 0x7fe0dfea7a60
calling cuInit
cuInit err: 802
time=2025-09-03T15:14:00.498Z level=INFO source=gpu.go:614 msg="Unable to load cudart library /usr/lib/x86_64-linux-gnu/libcuda.so.570.169: cuda driver library init failure: 802"

I just find it strange, because in the container, all my GPUs are visible, I can see them all with ls -l /dev/nvidia*, nvidia-smi works. It's just that whenever Ollama calls cuInit, I get that nasty 802 error.

Any thoughts? Like I said, I am new to all this and was kind of thrown to the wolves.

Relevant log output

OS

Docker

GPU

Nvidia

CPU

Other

Ollama version

0.11.8

Originally created by @connorgoodhue on GitHub (Sep 3, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/12173 ### What is the issue? I am new to Ollama, in fact, I am new to all of this. I've kind of been thrown blindly into setting up an "on-prem AI server" at my work. It is a Lambda Hyperplane server running 8 NVIDIA H100s. Anyway, I keep running into this error and I can't get it to go away. `Unable to load cudart library /usr/lib/x86_64-linux-gnu/libcuda.so.570.169: cuda driver library init failure: 802` The "AI server" is running Ubuntu Pro, and Ollama is running inside a Docker CE container (not Snap Docker). I have all the NVIDIA drivers installed at version 570.169 on the host, NVIDIA Container Toolkit is 1.17.8-1. Running CUDA 12.8. The CUDA drivers in the Ollama container are also 12.8. When I run nvidia-smi in the container it matches the versions shown on my host. I stopped trying to use the official Ollama Docker image, because I thought maybe I could just build off of an NVIDIA Cuda container image for 12.8 and then install Ollama on top of it. But that still didn't work, I get the same 802 error. Here is a more verbose version of the error. ``` time=2025-09-03T15:13:30.404Z level=INFO source=gpu.go:217 msg="looking for compatible GPUs" time=2025-09-03T15:13:30.458Z level=DEBUG source=gpu.go:98 msg="searching for GPU discovery libraries for NVIDIA" time=2025-09-03T15:13:30.458Z level=DEBUG source=gpu.go:503 msg="Searching for GPU library" name=libcuda.so* time=2025-09-03T15:13:30.458Z level=DEBUG source=gpu.go:527 msg="gpu library search" globs="[/usr/local/lib/ollama/libcuda.so* /usr/lib/x86_64-linux-gnu/libcuda.so* /usr/local/cuda/lib64/libcuda.so* /usr/local/cuda*/targets/*/lib/libcuda.so* /usr/lib/*-linux-gnu/nvidia/current/libcuda.so* /usr/lib/*-linux-gnu/libcuda.so* /usr/lib/wsl/lib/libcuda.so* /usr/lib/wsl/drivers/*/libcuda.so* /opt/cuda/lib*/libcuda.so* /usr/local/cuda/lib*/libcuda.so* /usr/lib*/libcuda.so* /usr/local/lib*/libcuda.so*]" time=2025-09-03T15:13:30.465Z level=DEBUG source=gpu.go:560 msg="discovered GPU libraries" paths=[/usr/lib/x86_64-linux-gnu/libcuda.so.570.169] initializing /usr/lib/x86_64-linux-gnu/libcuda.so.570.169 dlsym: cuInit - 0x7fe0dfe48a60 dlsym: cuDriverGetVersion - 0x7fe0dfe48a80 dlsym: cuDeviceGetCount - 0x7fe0dfe48ac0 dlsym: cuDeviceGet - 0x7fe0dfe48aa0 dlsym: cuDeviceGetAttribute - 0x7fe0dfe48ba0 dlsym: cuDeviceGetUuid - 0x7fe0dfe48b00 dlsym: cuDeviceGetName - 0x7fe0dfe48ae0 dlsym: cuCtxCreate_v3 - 0x7fe0dfe48d80 dlsym: cuMemGetInfo_v2 - 0x7fe0dfe69140 dlsym: cuCtxDestroy - 0x7fe0dfea7a60 calling cuInit cuInit err: 802 time=2025-09-03T15:14:00.498Z level=INFO source=gpu.go:614 msg="Unable to load cudart library /usr/lib/x86_64-linux-gnu/libcuda.so.570.169: cuda driver library init failure: 802" ``` I just find it strange, because in the container, all my GPUs are visible, I can see them all with ls -l /dev/nvidia*, nvidia-smi works. It's just that whenever Ollama calls cuInit, I get that nasty 802 error. Any thoughts? Like I said, I am new to all this and was kind of thrown to the wolves. ### Relevant log output ```shell ``` ### OS Docker ### GPU Nvidia ### CPU Other ### Ollama version 0.11.8
GiteaMirror added the bug label 2026-04-29 06:33:19 -05:00
GiteaMirror commented 2026-04-29 06:33:27 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Sep 3, 2025):

cudaErrorSystemNotReady = 802

  • This error indicates that the system is not yet ready to start any CUDA work. To continue using CUDA, verify the system configuration is in a valid state and all required driver daemons are actively running. More information about this error can be found in the system specific user guide.

See https://github.com/ollama/ollama/issues/9031, where a common fix is to install nvidia-fabricmanager-570.

<!-- gh-comment-id:3250099492 --> @rick-github commented on GitHub (Sep 3, 2025): [cudaErrorSystemNotReady](https://docs.nvidia.com/cuda/cuda-runtime-api/group__CUDART__TYPES.html#group__CUDART__TYPES_1gg3f51e3575c2178246db0a94a430e0038e942e4cbbd2bef6e92e293253f055613:~:text=cudaErrorSystemNotReady) = 802 - This error indicates that the system is not yet ready to start any CUDA work. To continue using CUDA, verify the system configuration is in a valid state and all required driver daemons are actively running. More information about this error can be found in the system specific user guide. See https://github.com/ollama/ollama/issues/9031, where a common fix is to install `nvidia-fabricmanager-570`.
GiteaMirror commented 2026-04-29 06:33:27 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 6, 2025):

Any luck with fixing as per #9031?

<!-- gh-comment-id:3369533826 --> @rick-github commented on GitHub (Oct 6, 2025): Any luck with fixing as per #9031?
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label bug
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#54609
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?