[GH-ISSUE #9592] No response after 2 minutes of idle time #68311

Closed
opened 2026-05-04 13:12:33 -05:00 by GiteaMirror · 1 comment
Owner

Originally created by @lynn158 on GitHub (Mar 8, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/9592

What is the issue?

Running the model and starting the conversation is normal, but after being idle for a few minutes, for example, 2 minutes of idle time, there will be no response. Ollama ps, please help me, I have changed several versions and it is still like this. The environment I use is Hyper-V to install Win11 virtual machine and set up vGPU. The test using the console and Cherry Studio is the same.
NAME ID SIZE

server.log

PROCESSOR UNTIL
qwq:latest cc1091b0e276 23 GB 100% GPU Stopping...
UNTIL will keep Stopping..., you can only manually end the ollama.exe process. (If you use Cherry Studio to manually end the process, it will continue to request until you manually end the ollama.exe process. At this time, ollama will re-run the model after it ends)
OLLAMA_KEEP_ALIVE uses the default 5m. After 5 minutes, UNTIL displays Stopping... You can only manually end the ollama.exe process

Relevant log output


OS

Windows

GPU

Nvidia

CPU

Intel

Ollama version

v0.5.13

Originally created by @lynn158 on GitHub (Mar 8, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/9592 ### What is the issue? Running the model and starting the conversation is normal, but after being idle for a few minutes, for example, 2 minutes of idle time, there will be no response. Ollama ps, please help me, I have changed several versions and it is still like this. The environment I use is Hyper-V to install Win11 virtual machine and set up vGPU. The test using the console and Cherry Studio is the same. NAME ID SIZE [server.log](https://github.com/user-attachments/files/19141409/server.log) PROCESSOR UNTIL qwq:latest cc1091b0e276 23 GB 100% GPU Stopping... UNTIL will keep Stopping..., you can only manually end the ollama.exe process. (If you use Cherry Studio to manually end the process, it will continue to request until you manually end the ollama.exe process. At this time, ollama will re-run the model after it ends) OLLAMA_KEEP_ALIVE uses the default 5m. After 5 minutes, UNTIL displays Stopping... You can only manually end the ollama.exe process ### Relevant log output ```shell ``` ### OS Windows ### GPU Nvidia ### CPU Intel ### Ollama version v0.5.13
GiteaMirror added the bug label 2026-05-04 13:12:33 -05:00
Author
Owner

@lynn158 commented on GitHub (Mar 8, 2025):

I think I know the reason. It may be related to my VPN software. After uninstalling the VPN software, it is now normal. Thanks again

<!-- gh-comment-id:2708196192 --> @lynn158 commented on GitHub (Mar 8, 2025): I think I know the reason. It may be related to my VPN software. After uninstalling the VPN software, it is now normal. Thanks again
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#68311