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 #11341] llama runner process has terminated: error:fault #53993

New Issue
Closed
opened 2026-04-29 05:04:21 -05:00 by GiteaMirror · 9 comments
GiteaMirror commented 2026-04-29 05:04:21 -05:00
Owner
Copy Link

Originally created by @ManishKhullar on GitHub (Jul 9, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/11341

What is the issue?

ollama run llama3.2-vision:11b
Error: llama runner process has terminated: error:fault

  • Got above error
  • For a windows system where the model is downloaded in G: drive
  • setup of ollama is done in G: drive
  • by default the CPU is being used despite having an Nvidia GeForce GTX 960 Graphics adapter

Below is the server side error:

Relevant log output

time=2025-07-09T14:23:42.280+05:30 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server not responding"
time=2025-07-09T14:23:42.639+05:30 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server error"
time=2025-07-09T14:23:42.682+05:30 level=ERROR source=server.go:464 msg="llama runner terminated" error="exit status 2"
time=2025-07-09T14:23:42.890+05:30 level=ERROR source=sched.go:489 msg="error loading llama server" error="llama runner process has terminated: error:fault"
[GIN] 2025/07/09 - 14:23:42 | 500 |    4.7371714s |       127.0.0.1 | POST     "/api/generate"
time=2025-07-09T14:23:47.925+05:30 level=WARN source=sched.go:687 msg="gpu VRAM usage didn't recover within timeout" seconds=5.0340531 runner.size="11.9 GiB" runner.vram="1.4 GiB" runner.parallel=1 runner.pid=12052 runner.model=G:\Manish\ollama\blobs\sha256-9999d473417a8e179d993498195be5f42cab963acc75f4a6b15d981e8b68abed
time=2025-07-09T14:23:48.175+05:30 level=WARN source=sched.go:687 msg="gpu VRAM usage didn't recover within timeout" seconds=5.2844668 runner.size="11.9 GiB" runner.vram="1.4 GiB" runner.parallel=1 runner.pid=12052 runner.model=G:\Manish\ollama\blobs\sha256-9999d473417a8e179d993498195be5f42cab963acc75f4a6b15d981e8b68abed
time=2025-07-09T14:23:48.425+05:30 level=WARN source=sched.go:687 msg="gpu VRAM usage didn't recover within timeout" seconds=5.5343092 runner.size="11.9 GiB" runner.vram="1.4 GiB" runner.parallel=1 runner.pid=12052 runner.model=G:\Manish\ollama\blobs\sha256-9999d473417a8e179d993498195be5f42cab963acc75f4a6b15d981e8b68abed

OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

0.9.6

Originally created by @ManishKhullar on GitHub (Jul 9, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/11341 ### What is the issue? ```ollama run llama3.2-vision:11b``` ```Error: llama runner process has terminated: error:fault``` - Got above error - For a windows system where the model is downloaded in G: drive - setup of ollama is done in G: drive - by default the CPU is being used despite having an Nvidia GeForce GTX 960 Graphics adapter Below is the server side error: ### Relevant log output ```shell time=2025-07-09T14:23:42.280+05:30 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server not responding" time=2025-07-09T14:23:42.639+05:30 level=INFO source=server.go:632 msg="waiting for server to become available" status="llm server error" time=2025-07-09T14:23:42.682+05:30 level=ERROR source=server.go:464 msg="llama runner terminated" error="exit status 2" time=2025-07-09T14:23:42.890+05:30 level=ERROR source=sched.go:489 msg="error loading llama server" error="llama runner process has terminated: error:fault" [GIN] 2025/07/09 - 14:23:42 | 500 | 4.7371714s | 127.0.0.1 | POST "/api/generate" time=2025-07-09T14:23:47.925+05:30 level=WARN source=sched.go:687 msg="gpu VRAM usage didn't recover within timeout" seconds=5.0340531 runner.size="11.9 GiB" runner.vram="1.4 GiB" runner.parallel=1 runner.pid=12052 runner.model=G:\Manish\ollama\blobs\sha256-9999d473417a8e179d993498195be5f42cab963acc75f4a6b15d981e8b68abed time=2025-07-09T14:23:48.175+05:30 level=WARN source=sched.go:687 msg="gpu VRAM usage didn't recover within timeout" seconds=5.2844668 runner.size="11.9 GiB" runner.vram="1.4 GiB" runner.parallel=1 runner.pid=12052 runner.model=G:\Manish\ollama\blobs\sha256-9999d473417a8e179d993498195be5f42cab963acc75f4a6b15d981e8b68abed time=2025-07-09T14:23:48.425+05:30 level=WARN source=sched.go:687 msg="gpu VRAM usage didn't recover within timeout" seconds=5.5343092 runner.size="11.9 GiB" runner.vram="1.4 GiB" runner.parallel=1 runner.pid=12052 runner.model=G:\Manish\ollama\blobs\sha256-9999d473417a8e179d993498195be5f42cab963acc75f4a6b15d981e8b68abed ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version 0.9.6
GiteaMirror added the bugneeds more info labels 2026-04-29 05:04:22 -05:00
GiteaMirror commented 2026-04-29 05:04:24 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 9, 2025):

Set OLLAMA_DEBUG=1 in the server environment and post the full log.

<!-- gh-comment-id:3051836471 --> @rick-github commented on GitHub (Jul 9, 2025): Set `OLLAMA_DEBUG=1` in the server environment and post the full log.
GiteaMirror commented 2026-04-29 05:04:24 -05:00
Author
Owner
Copy Link

@svaningelgem commented on GitHub (Jul 12, 2025):

I also encounter this issue (with another model)

Log attached: server.log

My previous version was 0.9.0, the one where I got this error on was 0.9.6

Going back to 0.9.5 made it work fine. (well, still figuring out how to generate images via ollama models, but I didn't get any error - and some irrelevant output) ;)

<!-- gh-comment-id:3064914915 --> @svaningelgem commented on GitHub (Jul 12, 2025): I also encounter this issue (with another model) Log attached: [server.log](https://github.com/user-attachments/files/21194759/server.log) My previous version was 0.9.0, the one where I got this error on was 0.9.6 Going back to 0.9.5 made it work fine. (well, still figuring out how to generate images via ollama models, but I didn't get any error - and some irrelevant output) ;)
GiteaMirror commented 2026-04-29 05:04:25 -05:00
Author
Owner
Copy Link

@mh67 commented on GitHub (Jul 12, 2025):

Hi all,
probably found solution (following description for Linux users):

  • Stop your ollama systemd (or similiar) service,
  • delete the directory /usr/lib/ollama,
  • untar the ollama package ollama-linux-amd64.tgz again by the help of command:
    tar -C /usr -xzvf ./ollama-linux-amd64.tgz
  • start your ollama service or start ollama manually according your environment...
  • ollama shoud work again without downgrade to older versions.

Why it works?
It seems there are serious changes in the /usr/lib/ollama directory. In case you only untar the new version from the .tgz file, several old libraries will remain on place and cause conflict... my hypothesis...

What does that imply? A full-featured package ( at least .deb, .rpm) is needed for ollama distribution... Quuestion for ollama authors or for somebody who wants to maintain these packages....
...or some very simple shell script may solve this...

<!-- gh-comment-id:3065125968 --> @mh67 commented on GitHub (Jul 12, 2025): Hi all, probably found solution (following description for Linux users): - Stop your ollama systemd (or similiar) service, - delete the directory /usr/lib/ollama, - untar the ollama package ollama-linux-amd64.tgz again by the help of command: tar -C /usr -xzvf ./ollama-linux-amd64.tgz - start your ollama service or start ollama manually according your environment... - ollama shoud work again without downgrade to older versions. Why it works? It seems there are serious changes in the /usr/lib/ollama directory. In case you only untar the new version from the .tgz file, several old libraries will remain on place and cause conflict... my hypothesis... What does that imply? A full-featured package ( at least .deb, .rpm) is needed for ollama distribution... Quuestion for ollama authors or for somebody who wants to maintain these packages.... ...or some very simple shell script may solve this...
GiteaMirror commented 2026-04-29 05:04:25 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 12, 2025):

In case you only untar the new version from the .tgz file

#11211

...or some very simple shell script may solve this...

https://ollama.com/download/linux

<!-- gh-comment-id:3065989848 --> @rick-github commented on GitHub (Jul 12, 2025): > In case you only untar the new version from the .tgz file #11211 > ...or some very simple shell script may solve this... https://ollama.com/download/linux
GiteaMirror commented 2026-04-29 05:04:25 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 12, 2025):

@svaningelgem You have a similar problem

load_backend: loaded ROCm backend from D:\Ollama\lib\ollama\ggml-hip.dll
load_backend: loaded ROCm backend from D:\Ollama\lib\ollama\rocm\ggml-hip.dll

ROCM backend is being loaded twice. Clear out D:\Ollama\lib and run the installer again.

<!-- gh-comment-id:3065992500 --> @rick-github commented on GitHub (Jul 12, 2025): @svaningelgem You have a similar problem ``` load_backend: loaded ROCm backend from D:\Ollama\lib\ollama\ggml-hip.dll load_backend: loaded ROCm backend from D:\Ollama\lib\ollama\rocm\ggml-hip.dll ``` ROCM backend is being loaded twice. Clear out D:\Ollama\lib and run the installer again.
GiteaMirror commented 2026-04-29 05:04:26 -05:00
Author
Owner
Copy Link

@svaningelgem commented on GitHub (Jul 17, 2025):

I can confirm that a removal (!keep the models!) & reinstallation made 0.9.6 work fine again.

Maybe an action point for the installer thus: see that all files are properly removed during removal (ie: first uninstall, then reinstall? Not just overwrite?)

<!-- gh-comment-id:3082326062 --> @svaningelgem commented on GitHub (Jul 17, 2025): I can confirm that a removal (!keep the models!) & reinstallation made 0.9.6 work fine again. Maybe an action point for the installer thus: see that all files are properly removed during removal (ie: first uninstall, then reinstall? Not just overwrite?)
GiteaMirror commented 2026-04-29 05:04:26 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 17, 2025):

I'm not a windows user, but my understanding is that the windows installer already does this, ie clears the directory before install, and that the problem arises when the files are moved after installation or are manually installed. Did you set a different installation directory during install?

<!-- gh-comment-id:3082345930 --> @rick-github commented on GitHub (Jul 17, 2025): I'm not a windows user, but my understanding is that the windows installer already does this, ie clears the directory before install, and that the problem arises when the files are moved after installation or are manually installed. Did you set a different installation directory during install?
GiteaMirror commented 2026-04-29 05:04:27 -05:00
Author
Owner
Copy Link

@svaningelgem commented on GitHub (Jul 17, 2025):

Yes, I'm installing on my D-drive (as C has no space left). I didn't do any manual manipulation though.
(To clarify: I have been always installing on my D-drive since the first installation, so it's not like I installed on C before, and now suddenly somewhere else)

<!-- gh-comment-id:3082351721 --> @svaningelgem commented on GitHub (Jul 17, 2025): Yes, I'm installing on my D-drive (as C has no space left). I didn't do any manual manipulation though. (To clarify: I have been always installing on my D-drive since the first installation, so it's not like I installed on C before, and now suddenly somewhere else)
GiteaMirror commented 2026-04-29 05:04:28 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jul 17, 2025):

Ok, thanks for the feedback.

<!-- gh-comment-id:3082371330 --> @rick-github commented on GitHub (Jul 17, 2025): Ok, thanks for the feedback.
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 needs more info
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#53993
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?