github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #8852] Ollama "No connection could be made because the target machine actively refused it." error only without GPU #52250

New Issue
Closed
opened 2026-04-28 22:39:23 -05:00 by GiteaMirror · 12 comments
GiteaMirror commented 2026-04-28 22:39:23 -05:00
Owner
Copy Link

Originally created by @Ltamann on GitHub (Feb 5, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/8852

What is the issue?

Tested on Windows Versions 0.55, 0.56, and 0.57:

  1. Configure NVIDIA Control Panel:

    • On Windows 11, open the NVIDIA Control Panel.
    • Navigate to "CUDA - GPUs" and set it to None (do not select any GPU).
    • Click Apply to save changes.

  2. Start the Ollama Server:

    • Open Command Prompt and run:
      ollama serve
    • Output:
      time= level=INFO source=gpu.go:392 msg="no compatible GPUs were discovered"
      time= level=INFO source=types.go:131 msg="inference compute" id=0 library=cpu variant=avx2 compute="" driver=0.0 name="" total="255.9 GiB" available="243.0 GiB"

  3. Run a Model:

    • In Command Prompt, run any Ollama model:
      ollama run [model-name]
    • Error encountered:
      Error: Post "http://0.0.0.0:11435/api/show": dial tcp 0.0.0.0:11435: connectex: No connection could be made because the target machine actively refused it.

Observation:

  • Using a GPU, the inference works correctly without issues:

    time= level=INFO source=types.go:131 msg="inference compute" id=GPU-.. library=cuda variant=v12 compute=8.6 driver=12.7 name="NVIDIA GeForce RTX 3090" total="24.0 GiB" available="22.8 GiB"

  • Without the GPU, the Ollama server shuts down after a minute, during model downloads, or while chatting. Always saying:No connection could be made because the target machine actively refused it.

Question:
What is a reliable way to run Ollama only using CPU and RAM instead of a GPU?

OS

Windows

GPU

Nvidia

CPU

AMD

Ollama version

0.55, 0.56, 0.57

Originally created by @Ltamann on GitHub (Feb 5, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/8852 ### What is the issue? Tested on Windows Versions 0.55, 0.56, and 0.57: 1. Configure NVIDIA Control Panel: - On Windows 11, open the NVIDIA Control Panel. - Navigate to "CUDA - GPUs" and set it to **None** (do not select any GPU). - Click **Apply** to save changes. 2. Start the Ollama Server: - Open Command Prompt and run: ollama serve - Output: time= level=INFO source=gpu.go:392 msg="no compatible GPUs were discovered" time= level=INFO source=types.go:131 msg="inference compute" id=0 library=cpu variant=avx2 compute="" driver=0.0 name="" total="255.9 GiB" available="243.0 GiB" 3. Run a Model: - In Command Prompt, run any Ollama model: ollama run [model-name] - Error encountered: Error: Post "http://0.0.0.0:11435/api/show": dial tcp 0.0.0.0:11435: connectex: No connection could be made because the target machine actively refused it. Observation: - Using a GPU, the inference works correctly without issues: time= level=INFO source=types.go:131 msg="inference compute" id=GPU-.. library=cuda variant=v12 compute=8.6 driver=12.7 name="NVIDIA GeForce RTX 3090" total="24.0 GiB" available="22.8 GiB" - Without the GPU, the Ollama server shuts down after a minute, during model downloads, or while chatting. Always saying:No connection could be made because the target machine actively refused it. Question: What is a reliable way to run Ollama only using CPU and RAM instead of a GPU? ### OS Windows ### GPU Nvidia ### CPU AMD ### Ollama version 0.55, 0.56, 0.57
GiteaMirror added the bug label 2026-04-28 22:39:23 -05:00
GiteaMirror commented 2026-04-28 22:39:27 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 5, 2025):

Without GPU, the Ollama server shuts down after starting and attempting to chat using downloaded models.

Does it actually shut down? What do the server logs show?

<!-- gh-comment-id:2637360026 --> @rick-github commented on GitHub (Feb 5, 2025): > Without GPU, the Ollama server shuts down after starting and attempting to chat using downloaded models. Does it actually shut down? What do the [server logs](https://github.com/ollama/ollama/blob/main/docs/troubleshooting.md#how-to-troubleshoot-issues) show?
GiteaMirror commented 2026-04-28 22:39:28 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 5, 2025):

And why port 11435?

<!-- gh-comment-id:2637363678 --> @rick-github commented on GitHub (Feb 5, 2025): And why port 11435?
GiteaMirror commented 2026-04-28 22:39:29 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 5, 2025):

Excuse the follow on posting.

What is a reliable way to run Ollama only using CPU and RAM instead of a GPU?

https://github.com/ollama/ollama/issues/6950#issuecomment-2373663650

<!-- gh-comment-id:2637370565 --> @rick-github commented on GitHub (Feb 5, 2025): Excuse the follow on posting. > What is a reliable way to run Ollama only using CPU and RAM instead of a GPU? https://github.com/ollama/ollama/issues/6950#issuecomment-2373663650
GiteaMirror commented 2026-04-28 22:39:30 -05:00
Author
Owner
Copy Link

@Ltamann commented on GitHub (Feb 5, 2025):

I switched ports to check if the issue was port-related, but it's not. The Ollama server shuts down after about a minute, even if I don't interact with it.

it's easy to replicate — just disable the GPU for Ollama, and you'll see the issue downloading a model.

The app.log shows a crash:
time=2025-02-05T16:37:08.749+01:00 level=WARN source=server.go:163 msg="server crash 1 - exit code 3221225477 - respawning"

Meanwhile, the server.log doesn't display any error messages.
2025/02/05 16:37:09 routes.go:1187: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\Users\YLAB-Partner\.ollama\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]"
time=2025-02-05T16:37:09.319+01:00 level=INFO source=images.go:432 msg="total blobs: 0"
time=2025-02-05T16:37:09.319+01:00 level=INFO source=images.go:439 msg="total unused blobs removed: 0"
time=2025-02-05T16:37:09.319+01:00 level=INFO source=routes.go:1238 msg="Listening on [::]:11434 (version 0.5.5)"
time=2025-02-05T16:37:09.320+01:00 level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11_avx cuda_v12_avx rocm_avx]"
time=2025-02-05T16:37:09.320+01:00 level=INFO source=gpu.go:226 msg="looking for compatible GPUs"
time=2025-02-05T16:37:09.320+01:00 level=INFO source=gpu_windows.go:167 msg=packages count=1
time=2025-02-05T16:37:09.320+01:00 level=INFO source=gpu_windows.go:214 msg="" package=0 cores=64 efficiency=0 threads=128
time=2025-02-05T16:37:09.362+01:00 level=INFO source=gpu.go:620 msg="no nvidia devices detected by library C:\Windows\system32\nvcuda.dll"
time=2025-02-05T16:37:09.496+01:00 level=INFO source=gpu.go:392 msg="no compatible GPUs were discovered"
time=2025-02-05T16:37:09.497+01:00 level=INFO source=types.go:131 msg="inference compute" id=0 library=cpu variant=avx2 compute="" driver=0.0 name="" total="255.9 GiB" available="242.5 GiB"

<!-- gh-comment-id:2637574851 --> @Ltamann commented on GitHub (Feb 5, 2025): I switched ports to check if the issue was port-related, but it's not. The Ollama server shuts down after about a minute, even if I don't interact with it. it's easy to replicate — just disable the GPU for Ollama, and you'll see the issue downloading a model. The app.log shows a crash: time=2025-02-05T16:37:08.749+01:00 level=WARN source=server.go:163 msg="server crash 1 - exit code 3221225477 - respawning" Meanwhile, the server.log doesn't display any error messages. 2025/02/05 16:37:09 routes.go:1187: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL: HIP_VISIBLE_DEVICES: HSA_OVERRIDE_GFX_VERSION: HTTPS_PROXY: HTTP_PROXY: NO_PROXY: OLLAMA_DEBUG:false OLLAMA_FLASH_ATTENTION:false OLLAMA_GPU_OVERHEAD:0 OLLAMA_HOST:http://0.0.0.0:11434 OLLAMA_INTEL_GPU:false OLLAMA_KEEP_ALIVE:5m0s OLLAMA_KV_CACHE_TYPE: OLLAMA_LLM_LIBRARY: OLLAMA_LOAD_TIMEOUT:5m0s OLLAMA_MAX_LOADED_MODELS:0 OLLAMA_MAX_QUEUE:512 OLLAMA_MODELS:C:\\Users\\YLAB-Partner\\.ollama\\models OLLAMA_MULTIUSER_CACHE:false OLLAMA_NOHISTORY:false OLLAMA_NOPRUNE:false OLLAMA_NUM_PARALLEL:0 OLLAMA_ORIGINS:[http://localhost https://localhost http://localhost:* https://localhost:* http://127.0.0.1 https://127.0.0.1 http://127.0.0.1:* https://127.0.0.1:* http://0.0.0.0 https://0.0.0.0 http://0.0.0.0:* https://0.0.0.0:* app://* file://* tauri://* vscode-webview://*] OLLAMA_SCHED_SPREAD:false ROCR_VISIBLE_DEVICES:]" time=2025-02-05T16:37:09.319+01:00 level=INFO source=images.go:432 msg="total blobs: 0" time=2025-02-05T16:37:09.319+01:00 level=INFO source=images.go:439 msg="total unused blobs removed: 0" time=2025-02-05T16:37:09.319+01:00 level=INFO source=routes.go:1238 msg="Listening on [::]:11434 (version 0.5.5)" time=2025-02-05T16:37:09.320+01:00 level=INFO source=routes.go:1267 msg="Dynamic LLM libraries" runners="[cpu cpu_avx cpu_avx2 cuda_v11_avx cuda_v12_avx rocm_avx]" time=2025-02-05T16:37:09.320+01:00 level=INFO source=gpu.go:226 msg="looking for compatible GPUs" time=2025-02-05T16:37:09.320+01:00 level=INFO source=gpu_windows.go:167 msg=packages count=1 time=2025-02-05T16:37:09.320+01:00 level=INFO source=gpu_windows.go:214 msg="" package=0 cores=64 efficiency=0 threads=128 time=2025-02-05T16:37:09.362+01:00 level=INFO source=gpu.go:620 msg="no nvidia devices detected by library C:\\Windows\\system32\\nvcuda.dll" time=2025-02-05T16:37:09.496+01:00 level=INFO source=gpu.go:392 msg="no compatible GPUs were discovered" time=2025-02-05T16:37:09.497+01:00 level=INFO source=types.go:131 msg="inference compute" id=0 library=cpu variant=avx2 compute="" driver=0.0 name="" total="255.9 GiB" available="242.5 GiB"
GiteaMirror commented 2026-04-28 22:39:30 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 5, 2025):

Try setting OLLAMA_DEBUG=1 in the server environment.

<!-- gh-comment-id:2637578763 --> @rick-github commented on GitHub (Feb 5, 2025): Try setting `OLLAMA_DEBUG=1` in the server environment.
GiteaMirror commented 2026-04-28 22:39:31 -05:00
Author
Owner
Copy Link

@sabbirsam commented on GitHub (Feb 5, 2025):

$ ollama run llama3.2-vision
pulling manifest
pulling manifest
pulling 11f274007f09... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 6.0 GB
pulling ece5e659647a... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 1.9 GB
pulling 715415638c9c... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 269 B
pulling 0b4284c1f870... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 7.7 KB
Error: Post "http://127.0.0.1:11434/api/show": read tcp 127.0.0.1:14014->127.0.0.1:11434: wsarecv: An existing connection was forcibly closed by the remote host.

After completing it shows "forcibly closed by the remote host"

<!-- gh-comment-id:2637597704 --> @sabbirsam commented on GitHub (Feb 5, 2025): $ ollama run llama3.2-vision pulling manifest pulling manifest pulling 11f274007f09... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 6.0 GB pulling ece5e659647a... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 1.9 GB pulling 715415638c9c... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 269 B pulling 0b4284c1f870... 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 7.7 KB Error: Post "http://127.0.0.1:11434/api/show": read tcp 127.0.0.1:14014->127.0.0.1:11434: wsarecv: An existing connection was forcibly closed by the remote host. After completing it shows "forcibly closed by the remote host"
GiteaMirror commented 2026-04-28 22:39:31 -05:00
Author
Owner
Copy Link

@Ltamann commented on GitHub (Feb 5, 2025):

Ollama debug : The last message before the server shuts down:

panic: runtime error: index out of range [0] with length 0

goroutine 83 [running]:
github.com/ollama/ollama/server.(*blobDownload).Prepare(0xc00033caf0, {0x7ff7b108ba80, 0xc0005a2be0}, 0xc000628480, 0xc0006139c0)
github.com/ollama/ollama/server/download.go:175 +0x539
github.com/ollama/ollama/server.downloadBlob({0x7ff7b108ba80, 0xc0005a2be0}, {{{0x7ff7b0ecb470, 0x5}, {0x7ff7b0ee02a7, 0x12}, {0x7ff7b0ed3ec3, 0x7}, {0xc000588d90, 0x7}, ...}, ...})
github.com/ollama/ollama/server/download.go:489 +0x4da
github.com/ollama/ollama/server.PullModel({0x7ff7b108ba80, 0xc0005a2be0}, {0xc000588d90, 0xe}, 0xc0006139c0, 0xc000048ee0)
github.com/ollama/ollama/server/images.go:564 +0x771
github.com/ollama/ollama/server.(*Server).PullHandler.func1()
github.com/ollama/ollama/server/routes.go:594 +0x197
created by github.com/ollama/ollama/server.(*Server).PullHandler in goroutine 49
github.com/ollama/ollama/server/routes.go:581 +0x691

<!-- gh-comment-id:2637598539 --> @Ltamann commented on GitHub (Feb 5, 2025): Ollama debug : The last message before the server shuts down: panic: runtime error: index out of range [0] with length 0 goroutine 83 [running]: github.com/ollama/ollama/server.(*blobDownload).Prepare(0xc00033caf0, {0x7ff7b108ba80, 0xc0005a2be0}, 0xc000628480, 0xc0006139c0) github.com/ollama/ollama/server/download.go:175 +0x539 github.com/ollama/ollama/server.downloadBlob({0x7ff7b108ba80, 0xc0005a2be0}, {{{0x7ff7b0ecb470, 0x5}, {0x7ff7b0ee02a7, 0x12}, {0x7ff7b0ed3ec3, 0x7}, {0xc000588d90, 0x7}, ...}, ...}) github.com/ollama/ollama/server/download.go:489 +0x4da github.com/ollama/ollama/server.PullModel({0x7ff7b108ba80, 0xc0005a2be0}, {0xc000588d90, 0xe}, 0xc0006139c0, 0xc000048ee0) github.com/ollama/ollama/server/images.go:564 +0x771 github.com/ollama/ollama/server.(*Server).PullHandler.func1() github.com/ollama/ollama/server/routes.go:594 +0x197 created by github.com/ollama/ollama/server.(*Server).PullHandler in goroutine 49 github.com/ollama/ollama/server/routes.go:581 +0x691
GiteaMirror commented 2026-04-28 22:39:33 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 5, 2025):

https://github.com/ollama/ollama/issues/8784

<!-- gh-comment-id:2637605849 --> @rick-github commented on GitHub (Feb 5, 2025): https://github.com/ollama/ollama/issues/8784
GiteaMirror commented 2026-04-28 22:39:33 -05:00
Author
Owner
Copy Link

@Ltamann commented on GitHub (Feb 5, 2025):

#8784

it’s only an issue when the GPU is disabled—everything works perfectly when the GPU is enabled, so I don’t think it’s the same problem. Also, the server doesn’t just shut down during downloads; it can shut down even when idle after waiting for a while.

<!-- gh-comment-id:2637861985 --> @Ltamann commented on GitHub (Feb 5, 2025): > [#8784](https://github.com/ollama/ollama/issues/8784) it’s only an issue when the GPU is disabled—everything works perfectly when the GPU is enabled, so I don’t think it’s the same problem. Also, the server doesn’t just shut down during downloads; it can shut down even when idle after waiting for a while.
GiteaMirror commented 2026-04-28 22:39:34 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Feb 5, 2025):

panic: runtime error: index out of range [0] with length 0

This error, at least, is #8784. If you have other logs, that may shed more light.

<!-- gh-comment-id:2637939851 --> @rick-github commented on GitHub (Feb 5, 2025): > panic: runtime error: index out of range [0] with length 0 This error, at least, is #8784. If you have other logs, that may shed more light.
GiteaMirror commented 2026-04-28 22:39:37 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 26, 2025):

The Ollama server shuts down after about a minute, even if I don't interact with it.
it's easy to replicate — just disable the GPU for Ollama, and you'll see the issue downloading a model.

Likely this: #9836

<!-- gh-comment-id:2754217247 --> @rick-github commented on GitHub (Mar 26, 2025): > The Ollama server shuts down after about a minute, even if I don't interact with it. > it's easy to replicate — just disable the GPU for Ollama, and you'll see the issue downloading a model. Likely this: #9836
GiteaMirror commented 2026-04-28 22:39:40 -05:00
Author
Owner
Copy Link

@dirkbrnd commented on GitHub (May 2, 2025):

This should fix it!
https://github.com/agno-agi/agno/pull/3057

I will release asap

<!-- gh-comment-id:2847264955 --> @dirkbrnd commented on GitHub (May 2, 2025): This should fix it! https://github.com/agno-agi/agno/pull/3057 I will release asap
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#52250
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?