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 #1994] Ollama requests hangs after about 20 requests and needs to be restarted #26909

New Issue
Closed
opened 2026-04-22 03:38:41 -05:00 by GiteaMirror · 5 comments
GiteaMirror commented 2026-04-22 03:38:41 -05:00
Owner
Copy Link

Originally created by @Shajan on GitHub (Jan 14, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/1994

Originally assigned to: @BruceMacD on GitHub.

Request hangs after about 20 requests.
Ollama version : 0.1.20, Linux with T4 GPU as well as Mac M2.
All subsequent api/generate request hangs for all models. The only way to resume is to restart ollama sudo systemctl restart ollama.

Repro

import requests

def query(session):
    url = "http://localhost:11434/api/generate"
    data = {
        "model": "llama2:7b",
        "prompt": "Why is the sky blue?",
        "stream": False,
    }

    with requests.post(url, json=data) as response: # Hangs about every 20 requests
        if response.ok:
            return response.text
        else:
            print(response)
            return None

def main():
    total = 0
    errors = 0

    with requests.Session() as session:
        for _ in range(100):
            total += 1
            r = query(session)
            if r is None:
                errors += 1
            success_rate = 100*((total - errors)/total)
            print(f"{total=} {errors=} {success_rate=:.2f}")

if __name__ == "__main__":
    main()
Originally created by @Shajan on GitHub (Jan 14, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/1994 Originally assigned to: @BruceMacD on GitHub. Request hangs after about 20 requests. Ollama version : 0.1.20, Linux with T4 GPU as well as Mac M2. All subsequent `api/generate` request hangs for all models. The only way to resume is to restart ollama `sudo systemctl restart ollama`. Repro ```python import requests def query(session): url = "http://localhost:11434/api/generate" data = { "model": "llama2:7b", "prompt": "Why is the sky blue?", "stream": False, } with requests.post(url, json=data) as response: # Hangs about every 20 requests if response.ok: return response.text else: print(response) return None def main(): total = 0 errors = 0 with requests.Session() as session: for _ in range(100): total += 1 r = query(session) if r is None: errors += 1 success_rate = 100*((total - errors)/total) print(f"{total=} {errors=} {success_rate=:.2f}") if __name__ == "__main__": main() ```
GiteaMirror commented 2026-04-22 03:38:42 -05:00
Author
Owner
Copy Link

@Shajan commented on GitHub (Jan 14, 2024):

Issue #1910 appears to be related. This issue appears to be different (unrelated to format='json').

<!-- gh-comment-id:1891066078 --> @Shajan commented on GitHub (Jan 14, 2024): Issue #1910 appears to be related. This issue appears to be different (unrelated to format='json').
GiteaMirror commented 2026-04-22 03:38:43 -05:00
Author
Owner
Copy Link

@Shajan commented on GitHub (Jan 15, 2024):

Same issue on versions 0.1.18, 0.1.19 (tested on linux)
Works fine on version 0.1.13 (tested for 1000 requests on linux)

<!-- gh-comment-id:1892640315 --> @Shajan commented on GitHub (Jan 15, 2024): Same issue on versions `0.1.18`, `0.1.19` (tested on linux) Works fine on version `0.1.13` (tested for 1000 requests on linux)
GiteaMirror commented 2026-04-22 03:38:43 -05:00
Author
Owner
Copy Link

@johanngerberding commented on GitHub (Jan 16, 2024):

Have the same issue with a gguf mistral model on a RTX6000 Quadro GPU on linux after 20-30 requests . Tested 0.1.13and 0.1.20.

<!-- gh-comment-id:1893623708 --> @johanngerberding commented on GitHub (Jan 16, 2024): Have the same issue with a gguf mistral model on a RTX6000 Quadro GPU on linux after 20-30 requests . Tested `0.1.13`and `0.1.20`.
GiteaMirror commented 2026-04-22 03:38:44 -05:00
Author
Owner
Copy Link

@BruceMacD commented on GitHub (Jan 16, 2024):

Thanks for the script in the report, I've reproduced this and found what is causing the issue. Working on getting to the root cause now.

<!-- gh-comment-id:1893954518 --> @BruceMacD commented on GitHub (Jan 16, 2024): Thanks for the script in the report, I've reproduced this and found what is causing the issue. Working on getting to the root cause now.
GiteaMirror commented 2026-04-22 03:38:45 -05:00
Author
Owner
Copy Link

@BruceMacD commented on GitHub (Jan 16, 2024):

We have a mitigation in for the next release by disabling prompt-caching: #2018

I'll follow up on why prompt-caching causes this in #2023

Thanks to everyone for the reports.

<!-- gh-comment-id:1894521948 --> @BruceMacD commented on GitHub (Jan 16, 2024): We have a mitigation in for the next release by disabling prompt-caching: #2018 I'll follow up on why prompt-caching causes this in #2023 Thanks to everyone for the reports.
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
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#26909
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?