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 #14757] Ollama is crashing on Ubuntu 25.10 for Claude Code tasks #35299

New Issue
Closed
opened 2026-04-22 19:42:28 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 19:42:28 -05:00
Owner
Copy Link

Originally created by @alex-ramanau on GitHub (Mar 10, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/14757

What is the issue?

When I run Claude Code using qwen3-coder-next on Ollama, I'm getting periodical crashes of Ollama servce(+ (core dumps).
It's happening after some time of GPU load, may happen after 10 minutes or 1 hour of heavy load. I'm trying mostly coding tasks, e.g. adding Prometheus metrics into small golang web service.

ollama.log

Relevant log output

OS details:

$ lsb_release -a                                 
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 25.10
Release:        25.10
Codename:       questing



CPU

╰─$ cat /proc/cpuinfo|grep AMD |grep model|head -n1
model name      : AMD RYZEN AI MAX+ 395 w/ Radeon 8060S


Claude Code config:

export ANTHROPIC_AUTH_TOKEN=ollama
export ANTHROPIC_API_KEY=""
export ANTHROPIC_BASE_URL=http://10.0.2.2:11434
export OLLAMA_MODEL=qwen3-coder-next
claude  --model qwen3-coder-next

OS

Linux

GPU

AMD

CPU

AMD

Ollama version

0.17.0

Originally created by @alex-ramanau on GitHub (Mar 10, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/14757 ### What is the issue? When I run Claude Code using `qwen3-coder-next` on Ollama, I'm getting periodical crashes of Ollama servce(+ (core dumps). It's happening after some time of GPU load, may happen after 10 minutes or 1 hour of heavy load. I'm trying mostly coding tasks, e.g. adding Prometheus metrics into small golang web service. [ollama.log](https://github.com/user-attachments/files/25861752/ollama.log) ### Relevant log output ```shell OS details: $ lsb_release -a No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 25.10 Release: 25.10 Codename: questing CPU ╰─$ cat /proc/cpuinfo|grep AMD |grep model|head -n1 model name : AMD RYZEN AI MAX+ 395 w/ Radeon 8060S Claude Code config: export ANTHROPIC_AUTH_TOKEN=ollama export ANTHROPIC_API_KEY="" export ANTHROPIC_BASE_URL=http://10.0.2.2:11434 export OLLAMA_MODEL=qwen3-coder-next claude --model qwen3-coder-next ``` ### OS Linux ### GPU AMD ### CPU AMD ### Ollama version 0.17.0
GiteaMirror added the bug label 2026-04-22 19:42:28 -05:00
GiteaMirror commented 2026-04-22 19:42:30 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 10, 2026):

сак 10 06:58:50 eniac-X2 ollama[75163]: HW Exception by GPU node-1 (Agent handle: 0x7dfe546f0130) reason :GPU Hang

This is a problem with the amdgpu kernel driver, it's been reported a few times and choosing a different kernel seems to help a bit, but it's not clear which kernel is best. Try switching to using Vulkan and see if the problem persists.

<!-- gh-comment-id:4030930289 --> @rick-github commented on GitHub (Mar 10, 2026): ``` сак 10 06:58:50 eniac-X2 ollama[75163]: HW Exception by GPU node-1 (Agent handle: 0x7dfe546f0130) reason :GPU Hang ``` This is a problem with the amdgpu kernel driver, it's been reported a few times and choosing a different kernel seems to help a bit, but it's not clear which kernel is best. Try switching to using [Vulkan](https://docs.ollama.com/gpu#vulkan-gpu-support) and see if the problem persists.
GiteaMirror commented 2026-04-22 19:42:30 -05:00
Author
Owner
Copy Link

@chejh-amd commented on GitHub (Mar 18, 2026):

A few additional things that might help narrow this down:

  1. Kernel version differences matter a lot on Ryzen‑AI / APU hardware.
    Several users have reported that 6.8 / 6.9 / 6.11 behave differently with the 8060S iGPU.
    If you’re on Ubuntu 25.10, testing an OEM kernel or LTS‑derived kernel can improve stability.
  2. Vulkan backend is often more stable than ROCm on Ryzen‑AI devices.
    Many APU systems avoid GPU hangs entirely when running: OLLAMA_VULKAN=1 ollama run
<!-- gh-comment-id:4079587261 --> @chejh-amd commented on GitHub (Mar 18, 2026): A few additional things that might help narrow this down: 1. Kernel version differences matter a lot on Ryzen‑AI / APU hardware. Several users have reported that 6.8 / 6.9 / 6.11 behave differently with the 8060S iGPU. If you’re on Ubuntu 25.10, testing an OEM kernel or LTS‑derived kernel can improve stability. 2. Vulkan backend is often more stable than ROCm on Ryzen‑AI devices. Many APU systems avoid GPU hangs entirely when running: `OLLAMA_VULKAN=1 ollama run`
GiteaMirror commented 2026-04-22 19:42:32 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Mar 20, 2026):

AMD recommends linux kernel 6.18.4 or newer for 8060S support.

<!-- gh-comment-id:4097477255 --> @rick-github commented on GitHub (Mar 20, 2026): AMD [recommends](https://rocm.docs.amd.com/en/latest/how-to/system-optimization/strixhalo.html#required-kernel-version) linux kernel 6.18.4 or newer for 8060S support.
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#35299
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?