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 #1266] Add a stop/restart command #648

New Issue
Closed
opened 2026-04-12 10:20:27 -05:00 by GiteaMirror · 14 comments
GiteaMirror commented 2026-04-12 10:20:27 -05:00
Owner
Copy Link

Originally created by @davlgd on GitHub (Nov 24, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1266

When I setup/launch ollama the manual way, I can launch the server with serve command but don't have a easy way to stop/restart it (so I need to kill the process). It would be great to have dedicated command for theses actions.

Originally created by @davlgd on GitHub (Nov 24, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1266 When I setup/launch `ollama` the manual way, I can launch the server with `serve` command but don't have a easy way to stop/restart it (so I need to kill the process). It would be great to have dedicated command for theses actions.
GiteaMirror added the feature request label 2026-04-12 10:20:27 -05:00
GiteaMirror commented 2026-04-12 10:20:28 -05:00
Author
Owner
Copy Link

@robbiemu commented on GitHub (Nov 25, 2023):

would love to see currently running models, and the url, in the menu bar icon (on Mac, that icon only offers "Quit" even once a model is running).

<!-- gh-comment-id:1826439818 --> @robbiemu commented on GitHub (Nov 25, 2023): would love to see currently running models, and the url, in the menu bar icon (on Mac, that icon only offers "Quit" even once a model is running).
GiteaMirror commented 2026-04-12 10:20:29 -05:00
Author
Owner
Copy Link

@technovangelist commented on GitHub (Dec 4, 2023):

Can you talk about why you want to restart the server?

and @robbiemu this seems to be an unrelated comment. But there is only one model at a time running. Whatever model is on the command line ollama run <model> is the one currently loaded.

<!-- gh-comment-id:1839753586 --> @technovangelist commented on GitHub (Dec 4, 2023): Can you talk about why you want to restart the server? and @robbiemu this seems to be an unrelated comment. But there is only one model at a time running. Whatever model is on the command line `ollama run <model>` is the one currently loaded.
GiteaMirror commented 2026-04-12 10:20:29 -05:00
Author
Owner
Copy Link

@robbiemu commented on GitHub (Dec 4, 2023):

@technovangelist you're right. sorry, I thought he meant individual models (I misread the OP). I was suggesting: if you are going to add cli to stop models (thank might ahve been started by API), you could also maybe add it to the gui

<!-- gh-comment-id:1839760538 --> @robbiemu commented on GitHub (Dec 4, 2023): @technovangelist you're right. sorry, I thought he meant individual models (I misread the OP). I was suggesting: if you are going to add cli to stop models (thank might ahve been started by API), you could also maybe add it to the gui
GiteaMirror commented 2026-04-12 10:20:29 -05:00
Author
Owner
Copy Link

@dudiao commented on GitHub (Dec 21, 2023):

I just wanted to stop the api server, for now I can only use the kill command

<!-- gh-comment-id:1866028280 --> @dudiao commented on GitHub (Dec 21, 2023): I just wanted to stop the api server, for now I can only use the kill command
GiteaMirror commented 2026-04-12 10:20:29 -05:00
Author
Owner
Copy Link

@mtrin commented on GitHub (Dec 29, 2023):

just like docker container stop exists, there should be some command to gracefully stop
for example I want it to pick up a new OLLAMA_MODELS env variable

<!-- gh-comment-id:1871965205 --> @mtrin commented on GitHub (Dec 29, 2023): just like docker container stop exists, there should be some command to gracefully stop for example I want it to pick up a new OLLAMA_MODELS env variable
GiteaMirror commented 2026-04-12 10:20:30 -05:00
Author
Owner
Copy Link

@technovangelist commented on GitHub (Jan 3, 2024):

On Mac, the way to stop Ollama is to click the menu bar icon and choose Quit Ollama. On Linux run sudo systemctl stop ollama

<!-- gh-comment-id:1875703294 --> @technovangelist commented on GitHub (Jan 3, 2024): On Mac, the way to stop Ollama is to click the menu bar icon and choose `Quit Ollama`. On Linux run `sudo systemctl stop ollama`
GiteaMirror commented 2026-04-12 10:20:30 -05:00
Author
Owner
Copy Link

@adrian-valente commented on GitHub (Feb 9, 2024):

I also think such an addition could be useful to allow non-root users to be able to free up GPU resources when the service is not needed, please tell me if there's a way to do this I am missing.

<!-- gh-comment-id:1935844912 --> @adrian-valente commented on GitHub (Feb 9, 2024): I also think such an addition could be useful to allow non-root users to be able to free up GPU resources when the service is not needed, please tell me if there's a way to do this I am missing.
GiteaMirror commented 2026-04-12 10:20:30 -05:00
Author
Owner
Copy Link

@chenxi1228 commented on GitHub (Feb 10, 2024):

On Mac, the way to stop Ollama is to click the menu bar icon and choose Quit Ollama. On Linux run sudo systemctl stop ollama

I'm wondering if I'm not a sudoer, how could I stop Ollama, since it will always occupy around 500MB GPU memory on each GPU (4 in total).

<!-- gh-comment-id:1936821782 --> @chenxi1228 commented on GitHub (Feb 10, 2024): > On Mac, the way to stop Ollama is to click the menu bar icon and choose `Quit Ollama`. On Linux run `sudo systemctl stop ollama` I'm wondering if I'm not a sudoer, how could I stop Ollama, since it will always occupy around 500MB GPU memory on each GPU (4 in total).
GiteaMirror commented 2026-04-12 10:20:30 -05:00
Author
Owner
Copy Link

@timiil commented on GitHub (Feb 19, 2024):

To give us a http endpoint that stop / restart the ollama service is good, specailly sometimes the ollama just hang up for the VRAM not realse correctly.

<!-- gh-comment-id:1951756693 --> @timiil commented on GitHub (Feb 19, 2024): To give us a http endpoint that stop / restart the ollama service is good, specailly sometimes the ollama just hang up for the VRAM not realse correctly.
GiteaMirror commented 2026-04-12 10:20:31 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Feb 20, 2024):

Hi all – to unload a model you can use the new keep_alive api parameter set to 0 https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-keep-a-model-loaded-in-memory-or-make-it-unload-immediately

Regarding stopping the Ollama service – you can send it a regular signal message with ctrl+c or kill. Let me know if this doesn't solve the issue though!

<!-- gh-comment-id:1953338470 --> @jmorganca commented on GitHub (Feb 20, 2024): Hi all – to unload a model you can use the new `keep_alive` api parameter set to `0` https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-keep-a-model-loaded-in-memory-or-make-it-unload-immediately Regarding stopping the Ollama service – you can send it a regular signal message with `ctrl+c` or `kill`. Let me know if this doesn't solve the issue though!
GiteaMirror commented 2026-04-12 10:20:31 -05:00
Author
Owner
Copy Link

@chenxi1228 commented on GitHub (Feb 20, 2024):

Hi all – to unload a model you can use the new keep_alive api parameter set to 0 https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-keep-a-model-loaded-in-memory-or-make-it-unload-immediately

Regarding stopping the Ollama service – you can send it a regular signal message with ctrl+c or kill. Let me know if this doesn't solve the issue though!

This can release the memory used by the model. However, the problem of "it will always occupy around 500MB GPU memory on each GPU (4 in total)" still exists. It cannot be stopped totally.

<!-- gh-comment-id:1953533823 --> @chenxi1228 commented on GitHub (Feb 20, 2024): > Hi all – to unload a model you can use the new `keep_alive` api parameter set to `0` https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-keep-a-model-loaded-in-memory-or-make-it-unload-immediately > > Regarding stopping the Ollama service – you can send it a regular signal message with `ctrl+c` or `kill`. Let me know if this doesn't solve the issue though! This can release the memory used by the model. However, the problem of "it will always occupy around 500MB GPU memory on each GPU (4 in total)" still exists. It cannot be stopped totally.
GiteaMirror commented 2026-04-12 10:20:32 -05:00
Author
Owner
Copy Link

@davlgd commented on GitHub (Feb 21, 2024):

Hi all – to unload a model you can use the new keep_alive api parameter set to 0 https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-keep-a-model-loaded-in-memory-or-make-it-unload-immediately

Regarding stopping the Ollama service – you can send it a regular signal message with ctrl+c or kill. Let me know if this doesn't solve the issue though!

My initial point on this was that, if I launch/use ollama as a server, I don't have any way to act on it as I have with the GUI. I can ollama serve but I don't have a way to do the opposite. I effectively detect it in processes and kill it in such situation but it could be great have a way to just ask it to ollama stop or even ollama restart (as we do with the GUI after an update for example).

<!-- gh-comment-id:1956072210 --> @davlgd commented on GitHub (Feb 21, 2024): > Hi all – to unload a model you can use the new `keep_alive` api parameter set to `0` https://github.com/ollama/ollama/blob/main/docs/faq.md#how-do-i-keep-a-model-loaded-in-memory-or-make-it-unload-immediately > > Regarding stopping the Ollama service – you can send it a regular signal message with `ctrl+c` or `kill`. Let me know if this doesn't solve the issue though! My initial point on this was that, if I launch/use ollama as a server, I don't have any way to act on it as I have with the GUI. I can `ollama serve` but I don't have a way to do the opposite. I effectively detect it in processes and kill it in such situation but it could be great have a way to just ask it to `ollama stop` or even `ollama restart` (as we do with the GUI after an update for example).
GiteaMirror commented 2026-04-12 10:20:32 -05:00
Author
Owner
Copy Link

@maxgreco commented on GitHub (Mar 10, 2024):

Hi, maybe for Windows this workaround ca be useful -> register this program for restart
https://www.thewindowsclub.com/what-does-register-this-program-for-restart-do
I will try it myself for may forked version of https://github.com/technovangelist/airenamer (BTW thanks a lot!)
I will follow this conversation, bye

<!-- gh-comment-id:1987105371 --> @maxgreco commented on GitHub (Mar 10, 2024): Hi, maybe for Windows this workaround ca be useful -> register this program for restart https://www.thewindowsclub.com/what-does-register-this-program-for-restart-do I will try it myself for may forked version of https://github.com/technovangelist/airenamer (BTW thanks a lot!) I will follow this conversation, bye
GiteaMirror commented 2026-04-12 10:20:33 -05:00
Author
Owner
Copy Link

@sheecet commented on GitHub (Apr 8, 2024):

this issue is not resolved

<!-- gh-comment-id:2042950793 --> @sheecet commented on GitHub (Apr 8, 2024): this issue is not resolved
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 feature request
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#648
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?