OpenWebUI - 500: Ollama: 500, message='Internal Server Error', url='http://0.0.0.0:11434/api/chat'

[Vendo233]

just installed the Open WebUI LXC, downloaded deepseek and I get this error

500: Ollama: 500, message='Internal Server Error', url='http://0.0.0.0:11434/api/chat'

tried to reinstall, 127.0.0.1, localhost etc.. nothing

[zelo66]

I I had the same problem. Found an error in the logs.
It turned out that this model needs more ram, but the error appeared the same.

[Vendo233]

yes, after increasing RAM and CPUs it works .. kind of. it is very slow though :-)

[Vendo233]

journalctl -u ollama.service -n

[HackerJL]

What did you increase it to to function?

[m4xwe11o]

Ahoi, I constantly run into the same error:

Apr 07 19:23:13 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-07T19:23:13.284Z level=ERROR source=server.go:449 msg="llama runner terminated" error="exit status 2"
Apr 07 19:23:13 at-vie-tds-lab-ollama-srv01 ollama[6243]: [GIN] 2025/04/07 - 19:23:13 | 500 | 15m52s | 127.0.0.1 | POST "/api/chat"
Apr 07 19:28:18 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-07T19:28:18.442Z level=WARN source=sched.go:648 msg="gpu VRAM usage didn't recover within timeout" seconds=5.113911195 model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
Apr 07 19:28:18 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-07T19:28:18.693Z level=WARN source=sched.go:648 msg="gpu VRAM usage didn't recover within timeout" seconds=5.364348192 model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
Apr 07 19:28:18 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-07T19:28:18.943Z level=WARN source=sched.go:648 msg="gpu VRAM usage didn't recover within timeout" seconds=5.6147561150000005 model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
....
Apr 08 18:53:21 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-08T18:53:21.913Z level=ERROR source=server.go:449 msg="llama runner terminated" error="exit status 2"
Apr 08 18:53:22 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-08T18:53:22.100Z level=ERROR source=sched.go:457 msg="error loading llama server" error="llama runner process has terminated: CUDA error: operation not supported\n current device: 0, in function ggml_backend_cuda_device_get_memory at //ml/backend/ggml/ggml/src/ggml-cuda/ggml-cuda.cu:2898\n cudaMemGetInfo(free, total)\n//ml/backend/ggml/ggml/src/ggml-cuda/ggml-cuda.cu:73: CUDA error"
Apr 08 18:53:22 at-vie-tds-lab-ollama-srv01 ollama[6243]: [GIN] 2025/04/08 - 18:53:22 | 500 | 672.233005ms | 127.0.0.1 | POST "/api/chat"
Apr 08 18:53:27 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-08T18:53:27.208Z level=WARN source=sched.go:648 msg="gpu VRAM usage didn't recover within timeout" seconds=5.107402319 model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
Apr 08 18:53:27 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-08T18:53:27.457Z level=WARN source=sched.go:648 msg="gpu VRAM usage didn't recover within timeout" seconds=5.356868987 model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e
Apr 08 18:53:27 at-vie-tds-lab-ollama-srv01 ollama[6243]: time=2025-04-08T18:53:27.708Z level=WARN source=sched.go:648 msg="gpu VRAM usage didn't recover within timeout" seconds=5.60736396 model=/usr/share/ollama/.ollama/models/blobs/sha256-6e9f90f02bb3b39b59e81916e8cfce9deb45aeaeb9a54a5be4414486b907dc1e

after rebooting the service it runs again.
The vm I set up has the following specs

[MickLesk]

Why Here? Its an Ollama issue