github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 05:05:37 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #3341] When is ollama serve ready - for use in scripts? #2058

New Issue
Open
opened 2026-04-12 12:17:48 -05:00 by GiteaMirror · 9 comments
GiteaMirror commented 2026-04-12 12:17:48 -05:00
Owner
Copy Link

Originally created by @alexellis on GitHub (Mar 25, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/3341

What are you trying to do?

I want to start ollama serve in the background for automation purposes, and then be able to run something like ollama ready which would block until the serve has loaded.

I see this take up to 5 seconds with an Nvidia 3060.

How should we solve this?

ollama ready would be ideal or ollama serve --ready or similar CLI command. Having a REST endpoint could work, but would mean writing a lot of bash loops which can be fragile.

What is the impact of not solving this?

Having to use ollama serve & followed by sleep 5 and that not working some of the time.

Anything else?

I couldn't find anything in the docs that talks about this, but if there is a mechanism for this already, I'd be happy to try it out.

Here's a blog post showing ollama being automated with bash in GitHub Actions, with a GPU attached:

https://actuated.dev/blog/ollama-in-github-actions

Originally created by @alexellis on GitHub (Mar 25, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/3341 ### What are you trying to do? I want to start ollama serve in the background for automation purposes, and then be able to run something like `ollama ready` which would block until the serve has loaded. I see this take up to 5 seconds with an Nvidia 3060. ### How should we solve this? `ollama ready` would be ideal or `ollama serve --ready` or similar CLI command. Having a REST endpoint could work, but would mean writing a lot of bash loops which can be fragile. ### What is the impact of not solving this? Having to use `ollama serve &` followed by `sleep 5` and that not working some of the time. ### Anything else? I couldn't find anything in the docs that talks about this, but if there is a mechanism for this already, I'd be happy to try it out. Here's a blog post showing ollama being automated with bash in GitHub Actions, with a GPU attached: https://actuated.dev/blog/ollama-in-github-actions
GiteaMirror added the feature request label 2026-04-12 12:17:48 -05:00
GiteaMirror commented 2026-04-12 12:17:49 -05:00
Author
Owner
Copy Link

@Kreijstal commented on GitHub (Mar 25, 2024):

so a sync block of ollama serve

<!-- gh-comment-id:2017749635 --> @Kreijstal commented on GitHub (Mar 25, 2024): so a sync block of ollama serve
GiteaMirror commented 2026-04-12 12:17:50 -05:00
Author
Owner
Copy Link

@alexellis commented on GitHub (Mar 25, 2024):

How would that look in bash?

I can't see how we could run additional commands if we were permanently blocked on ollama serve?

Screenshot from 2024-03-25 11-08-55

The sleep is not deterministic so it's starting to generate misc errors.

<!-- gh-comment-id:2017754269 --> @alexellis commented on GitHub (Mar 25, 2024): How would that look in bash? I can't see how we could run additional commands if we were permanently blocked on ollama serve? ![Screenshot from 2024-03-25 11-08-55](https://github.com/ollama/ollama/assets/6358735/4006167d-aaba-4713-a8bb-2a0f7e010c7c) The sleep is not deterministic so it's starting to generate misc errors.
GiteaMirror commented 2026-04-12 12:17:50 -05:00
Author
Owner
Copy Link

@alexellis commented on GitHub (Mar 25, 2024):

That sounds like an awful lot of work and repetition in bash.

I would suggest that the ollama team should add this as a Go command or flag and it'd be a very helpful addition.

The ollama serve command should ideally have a --daemonise flag too.

@Kreijstal are you a maintainer/team member? I wasn't sure from your GH profile

<!-- gh-comment-id:2017814362 --> @alexellis commented on GitHub (Mar 25, 2024): That sounds like an awful lot of work and repetition in bash. I would suggest that the ollama team should add this as a Go command or flag and it'd be a very helpful addition. The `ollama serve` command should ideally have a `--daemonise` flag too. @Kreijstal are you a maintainer/team member? I wasn't sure from your GH profile
GiteaMirror commented 2026-04-12 12:17:51 -05:00
Author
Owner
Copy Link

@Kreijstal commented on GitHub (Mar 25, 2024):

That sounds like an awful lot of work and repetition in bash.

I would suggest that the ollama team should add this as a Go command or flag and it'd be a very helpful addition.

The ollama serve command should ideally have a --daemonise flag too.

@Kreijstal are you a maintainer/team member? I wasn't sure from your GH profile

no just user

<!-- gh-comment-id:2017834907 --> @Kreijstal commented on GitHub (Mar 25, 2024): > That sounds like an awful lot of work and repetition in bash. > > I would suggest that the ollama team should add this as a Go command or flag and it'd be a very helpful addition. > > The `ollama serve` command should ideally have a `--daemonise` flag too. > > @Kreijstal are you a maintainer/team member? I wasn't sure from your GH profile no just user
GiteaMirror commented 2026-04-12 12:17:51 -05:00
Author
Owner
Copy Link

@pdevine commented on GitHub (Mar 25, 2024):

Hey @alexellis! For the endpoint solution you mentioned, you should be able to curl localhost:11434 and it will only respond w/ a 200 if the server is up and running.

<!-- gh-comment-id:2018381884 --> @pdevine commented on GitHub (Mar 25, 2024): Hey @alexellis! For the endpoint solution you mentioned, you should be able to `curl localhost:11434` and it will only respond w/ a 200 if the server is up and running.
GiteaMirror commented 2026-04-12 12:17:52 -05:00
Author
Owner
Copy Link

@remy415 commented on GitHub (Mar 25, 2024):

@alexellis In order to run something as a "daemon", you need to have some kind of service manager in your OS. In Linux, which seems to be the environment you're asking about, this is commonly one of initd, sysv, or systemd. You also could technically use Docker or Kubernetes as a "daemon" in a way, but that is out of scope here.

Without some kind of service manager, the only way to run something as a daemon is to mybinary &, nohup mybinary, tmux, etc, so that it runs without capturing your current terminal.

Does your platform have something like systemd on it? If so, just reference the ollama install script to set it up as a service and go to town.

Additionally, did you see this part of the article you mentioned:
image

Did you try using the time curl -i localhost:11434 command referenced there as a blocker agent?

<!-- gh-comment-id:2018850144 --> @remy415 commented on GitHub (Mar 25, 2024): @alexellis In order to run something as a "daemon", you need to have some kind of service manager in your OS. In Linux, which seems to be the environment you're asking about, this is commonly one of `initd`, `sysv`, or `systemd`. You also could technically use Docker or Kubernetes as a "daemon" in a way, but that is out of scope here. Without some kind of service manager, the only way to run something as a daemon is to `mybinary &`, `nohup mybinary`, `tmux`, etc, so that it runs without capturing your current terminal. Does your platform have something like `systemd` on it? If so, just reference the ollama install script to set it up as a service and go to town. Additionally, did you see this part of the article you mentioned: ![image](https://github.com/ollama/ollama/assets/105550370/c1ee6d14-c814-4f54-9c03-ed9dac5150d1) Did you try using the `time curl -i localhost:11434` command referenced there as a blocker agent?
GiteaMirror commented 2026-04-12 12:17:52 -05:00
Author
Owner
Copy Link

@alexellis commented on GitHub (Mar 27, 2024):

The curl fails if it follows ollama serve, due to the port not being open immediately. A sleep of a few seconds gives enough time until the port is open.

<!-- gh-comment-id:2022683782 --> @alexellis commented on GitHub (Mar 27, 2024): The curl fails if it follows ollama serve, due to the port not being open immediately. A sleep of a few seconds gives enough time until the port is open.
GiteaMirror commented 2026-04-12 12:17:53 -05:00
Author
Owner
Copy Link

@alexellis commented on GitHub (Jun 5, 2024):

@royjhan did you have any further thoughts on the thread?

The curl fails if it follows ollama serve due to the port not being opened immediately.. a sleep is required which is liable to be too long or to fail.

<!-- gh-comment-id:2150951771 --> @alexellis commented on GitHub (Jun 5, 2024): @royjhan did you have any further thoughts on the thread? The curl fails if it follows `ollama serve` due to the port not being opened immediately.. a sleep is required which is liable to be too long or to fail.
GiteaMirror commented 2026-04-12 12:17:54 -05:00
Author
Owner
Copy Link

@codefromthecrypt commented on GitHub (Sep 19, 2024):

I'd suggest this path, which seems to work for me alright. If folks are keen maybe we can help close out this issue so folks can reserve energy?

      - name: Start Ollama
        run: |
            # Run the background, in a way that survives to the next step
            nohup ollama serve > ollama.log 2>&1 &
            # Block using the ready endpoint
            time curl --retry 5 --retry-connrefused --retry-delay 0 -sf http://localhost:11434
<!-- gh-comment-id:2360577772 --> @codefromthecrypt commented on GitHub (Sep 19, 2024): I'd suggest this path, which seems to work for me alright. If folks are keen maybe we can help close out this issue so folks can reserve energy? ```yaml - name: Start Ollama run: | # Run the background, in a way that survives to the next step nohup ollama serve > ollama.log 2>&1 & # Block using the ready endpoint time curl --retry 5 --retry-connrefused --retry-delay 0 -sf http://localhost:11434 ```
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#2058
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?