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 #15770] Issue 0.21.x mlx runner error 500 #56561

New Issue
Closed
opened 2026-04-29 11:01:25 -05:00 by GiteaMirror · 10 comments
GiteaMirror commented 2026-04-29 11:01:25 -05:00
Owner
Copy Link

Originally created by @Sko85ch on GitHub (Apr 23, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/15770

What is the issue?

Error: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90

Relevant log output

OS

macOS

GPU

No response

CPU

Apple

Ollama version

0.21.1

Originally created by @Sko85ch on GitHub (Apr 23, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/15770 ### What is the issue? Error: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90 ### Relevant log output ```shell ``` ### OS macOS ### GPU _No response_ ### CPU Apple ### Ollama version 0.21.1
GiteaMirror added the bug label 2026-04-29 11:01:25 -05:00
GiteaMirror commented 2026-04-29 11:01:34 -05:00
Author
Owner
Copy Link

@albihasani94 commented on GitHub (Apr 23, 2026):

I am running into the same issue trying to run either qwen3.6:35b-a3b-coding-nvfp4 or qwen3.6:27b-coding-nvfp4 on Apple M2 Pro with 16 GPU core count. If there was any other info to share to help maintainers, would be:

panic: mlx: There is no Stream(gpu, 1) in current thread. at .../mlx-c-0.6.0/mlx/c/transforms.cpp:73
        panic: mlx: There is no Stream(gpu, 1) in current thread. at .../mlx-c-0.6.0/mlx/c/transforms.cpp:15

Previously,

source=client.go:420 msg="starting mlx runner subprocess" model=qwen3.6:27b-coding-nvfp4
source=sched.go:561 msg="loaded runners" count=1
source=server.go:31 msg="MLX engine initialized" "MLX version"=0.31.2 device=gpu
...
source=client.go:147 msg="mlx runner is ready"
<!-- gh-comment-id:4304138897 --> @albihasani94 commented on GitHub (Apr 23, 2026): I am running into the same issue trying to run either `qwen3.6:35b-a3b-coding-nvfp4` or `qwen3.6:27b-coding-nvfp4` on Apple M2 Pro with 16 GPU core count. If there was any other info to share to help maintainers, would be: ``` panic: mlx: There is no Stream(gpu, 1) in current thread. at .../mlx-c-0.6.0/mlx/c/transforms.cpp:73 panic: mlx: There is no Stream(gpu, 1) in current thread. at .../mlx-c-0.6.0/mlx/c/transforms.cpp:15 ``` Previously, ``` source=client.go:420 msg="starting mlx runner subprocess" model=qwen3.6:27b-coding-nvfp4 source=sched.go:561 msg="loaded runners" count=1 source=server.go:31 msg="MLX engine initialized" "MLX version"=0.31.2 device=gpu ... source=client.go:147 msg="mlx runner is ready" ```
GiteaMirror commented 2026-04-29 11:01:37 -05:00
Author
Owner
Copy Link

@ohjakobsen commented on GitHub (Apr 23, 2026):

I'm seeing the same as well on an M4 Mac with gemma4:26b-nvfp4:

panic: mlx: There is no Stream(gpu, 2) in current thread. at /private/tmp/mlx-c-20260422-18059-3ipqfu/mlx-c-0.6.0/mlx/c/transforms.cpp:15

Same MLX-version

source=server.go:31 msg="MLX engine initialized" "MLX version"=0.31.2 device=gpu
<!-- gh-comment-id:4307375142 --> @ohjakobsen commented on GitHub (Apr 23, 2026): I'm seeing the same as well on an M4 Mac with `gemma4:26b-nvfp4`: ``` panic: mlx: There is no Stream(gpu, 2) in current thread. at /private/tmp/mlx-c-20260422-18059-3ipqfu/mlx-c-0.6.0/mlx/c/transforms.cpp:15 ``` Same MLX-version ``` source=server.go:31 msg="MLX engine initialized" "MLX version"=0.31.2 device=gpu ```
GiteaMirror commented 2026-04-29 11:01:39 -05:00
Author
Owner
Copy Link

@Sko85ch commented on GitHub (Apr 24, 2026):

Issue also in 0.21.2

<!-- gh-comment-id:4311975252 --> @Sko85ch commented on GitHub (Apr 24, 2026): Issue also in 0.21.2
GiteaMirror commented 2026-04-29 11:01:40 -05:00
Author
Owner
Copy Link

@StefVanG commented on GitHub (Apr 24, 2026):

Experienced the exact same error on Ollama v0.21.1 and 0.21.2 and qwen3.6:27b-mlx-bf16, qwen3.6:27b-coding-bf16, qwen3.6:35b-a3b-coding-bf16. qwen3-coder-next:q4_K_M, devstral-small-2:24b-instruct-2512-fp16 and devstral-2:123b-instruct-2512-q4_K_M have no problem. M5 Max with 40 Core GPU and 128GB Ram

<!-- gh-comment-id:4313481781 --> @StefVanG commented on GitHub (Apr 24, 2026): Experienced the exact same error on Ollama v0.21.1 and 0.21.2 and qwen3.6:27b-mlx-bf16, qwen3.6:27b-coding-bf16, qwen3.6:35b-a3b-coding-bf16. qwen3-coder-next:q4_K_M, devstral-small-2:24b-instruct-2512-fp16 and devstral-2:123b-instruct-2512-q4_K_M have no problem. M5 Max with 40 Core GPU and 128GB Ram
GiteaMirror commented 2026-04-29 11:01:41 -05:00
Author
Owner
Copy Link

@antoyavin commented on GitHub (Apr 24, 2026):

Same issue here on ollama 0.21.2:

ollama run qwen3.5:35b-a3b-coding-nvfp4
pulling manifest
pulling model: 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████████▏  21 GB
writing manifest
success
>>> hello
Error: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90

M5 with 10 Core GPU and 32GB RAM

<!-- gh-comment-id:4314505523 --> @antoyavin commented on GitHub (Apr 24, 2026): Same issue here on ollama 0.21.2: ```bash ollama run qwen3.5:35b-a3b-coding-nvfp4 pulling manifest pulling model: 100% ▕██████████████████████████████████████████████████████████████████████████████████████████████████████████████▏ 21 GB writing manifest success >>> hello Error: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90 ``` M5 with 10 Core GPU and 32GB RAM
GiteaMirror commented 2026-04-29 11:01:42 -05:00
Author
Owner
Copy Link

@oldfarth commented on GitHub (Apr 24, 2026):

Same error like antoyavin
M1 Max 64GB RAM

==> ollama: stable 0.21.2

==> mlx-c: stable 0.6.0

==> mlx: stable 0.31.2
==> Dependencies
Required (1): python@3.14
Recursive Runtime (9)
Dependents: 2
==> Requirements
Build: Xcode >= 15.0 (on macOS)
Required: arm64 architecture, macOS >= 14 (or Linux), macOS

Thank you antoyavin. https://github.com/ollama/ollama/pull/15793 solved the problem

<!-- gh-comment-id:4316378051 --> @oldfarth commented on GitHub (Apr 24, 2026): Same error like [antoyavin](https://github.com/antoyavin) M1 Max 64GB RAM ==> ollama: stable 0.21.2 ==> mlx-c: stable 0.6.0 ==> mlx: stable 0.31.2 ==> Dependencies Required (1): python@3.14 Recursive Runtime (9) Dependents: 2 ==> Requirements Build: Xcode >= 15.0 (on macOS) Required: arm64 architecture, macOS >= 14 (or Linux), macOS Thank you [antoyavin](https://github.com/antoyavin). https://github.com/ollama/ollama/pull/15793 solved the problem
GiteaMirror commented 2026-04-29 11:01:43 -05:00
Author
Owner
Copy Link

@IAmBrendanL commented on GitHub (Apr 24, 2026):

I'm also running into the same issue on an M5 pro with 20 GPU cores when trying to use qwen3.6:27b-coding-mxfp8
Error: error running model: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90

<!-- gh-comment-id:4316994878 --> @IAmBrendanL commented on GitHub (Apr 24, 2026): I'm also running into the same issue on an M5 pro with 20 GPU cores when trying to use qwen3.6:27b-coding-mxfp8 `Error: error running model: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90`
GiteaMirror commented 2026-04-29 11:01:44 -05:00
Author
Owner
Copy Link

@andreinknv commented on GitHub (Apr 25, 2026):

Here is fix for the issue https://github.com/ollama/ollama/pull/15793

<!-- gh-comment-id:4317560817 --> @andreinknv commented on GitHub (Apr 25, 2026): Here is fix for the issue https://github.com/ollama/ollama/pull/15793
GiteaMirror commented 2026-04-29 11:01:45 -05:00
Author
Owner
Copy Link

@2pl commented on GitHub (Apr 25, 2026):

Same issue here.

  • Apple M4 / Tahoe 26.2 / 32GB
  • ollama installed from brew:
    • ollama version is 0.21.2
    • mlx 0.31.2
ollama run  qwen3.6:27b-coding-nvfp4 --verbose <<< "Hi"

Gives

Error: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90

In ollama logs I have

panic: mlx: There is no Stream(gpu, 1) in current thread. at /private/tmp/mlx-c-20260422-18059-3ipqfu/mlx-c-0.6.0/mlx/c/transforms.cpp:73

did not try the fix pointed to by @andreinknv yet

<!-- gh-comment-id:4318510466 --> @2pl commented on GitHub (Apr 25, 2026): Same issue here. - Apple M4 / Tahoe 26.2 / 32GB - ollama installed from brew: - ollama version is 0.21.2 - mlx 0.31.2 ```shell ollama run qwen3.6:27b-coding-nvfp4 --verbose <<< "Hi" ``` Gives ```shell Error: 500 Internal Server Error: mlx runner failed: golang.org/x/sync@v0.17.0/errgroup/errgroup.go:78 +0x90 ``` In ollama logs I have > panic: mlx: There is no Stream(gpu, 1) in current thread. at /private/tmp/mlx-c-20260422-18059-3ipqfu/mlx-c-0.6.0/mlx/c/transforms.cpp:73 did not try the fix pointed to by @andreinknv yet
GiteaMirror commented 2026-04-29 11:01:46 -05:00
Author
Owner
Copy Link

@Sko85ch commented on GitHub (Apr 25, 2026):

Solved. thx!

<!-- gh-comment-id:4320285886 --> @Sko85ch commented on GitHub (Apr 25, 2026): Solved. thx!
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#56561
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?