github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-06 16:11:34 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #1158] max retries exceeded: unexpected EOF #587

New Issue
Closed
opened 2026-04-12 10:17:21 -05:00 by GiteaMirror · 15 comments
GiteaMirror commented 2026-04-12 10:17:21 -05:00
Owner
Copy Link

Originally created by @priamai on GitHub (Nov 16, 2023).
Original GitHub issue: https://github.com/ollama/ollama/issues/1158

Originally assigned to: @BruceMacD on GitHub.

Hi there,
I am not sure if this is related to your file service, but I am getting this connection drops out very often.

Screenshot from 2023-11-16 23-45-54

Maybe there is a way to throttle requests?

Originally created by @priamai on GitHub (Nov 16, 2023). Original GitHub issue: https://github.com/ollama/ollama/issues/1158 Originally assigned to: @BruceMacD on GitHub. Hi there, I am not sure if this is related to your file service, but I am getting this connection drops out very often. ![Screenshot from 2023-11-16 23-45-54](https://github.com/jmorganca/ollama/assets/57333254/d530f24e-af82-49d8-9435-0653922d1eec) Maybe there is a way to throttle requests?
GiteaMirror added the bug label 2026-04-12 10:17:21 -05:00
GiteaMirror commented 2026-04-12 10:17:24 -05:00
Author
Owner
Copy Link

@orlyandico commented on GitHub (Nov 17, 2023):

Just keep re-running the pull, it will resume from where it left off. I encounter it a lot because I only have a 200Mbps connection and had three PC's downloading models at one time... (it would be more efficient to download once and then SCP the blobs to the other PC's).

<!-- gh-comment-id:1815588408 --> @orlyandico commented on GitHub (Nov 17, 2023): Just keep re-running the pull, it will resume from where it left off. I encounter it a lot because I only have a 200Mbps connection and had three PC's downloading models at one time... (it would be more efficient to download once and then SCP the blobs to the other PC's).
GiteaMirror commented 2026-04-12 10:17:25 -05:00
Author
Owner
Copy Link

@igorschlum commented on GitHub (Nov 17, 2023):

I have this problem also. More often when my internet connection was poor. In some case I had to relaunch the pull more than 20 times. It will become an issue if a model is updated and if Ollama runs as a stand alone server. Is it possible to extend the number of retries?

<!-- gh-comment-id:1815728236 --> @igorschlum commented on GitHub (Nov 17, 2023): I have this problem also. More often when my internet connection was poor. In some case I had to relaunch the pull more than 20 times. It will become an issue if a model is updated and if Ollama runs as a stand alone server. Is it possible to extend the number of retries?
GiteaMirror commented 2026-04-12 10:17:25 -05:00
Author
Owner
Copy Link

@priamai commented on GitHub (Nov 17, 2023):

I have a fiber connection here with 300 Mbps, I believe the issue is where the file is downloaded from. What about putting those files on a more reliable service like AWS S3?

<!-- gh-comment-id:1815903150 --> @priamai commented on GitHub (Nov 17, 2023): I have a fiber connection here with 300 Mbps, I believe the issue is where the file is downloaded from. What about putting those files on a more reliable service like AWS S3?
GiteaMirror commented 2026-04-12 10:17:26 -05:00
Author
Owner
Copy Link

@igorschlum commented on GitHub (Nov 17, 2023):

You raise a good point, @priamai. Having a better bandwidth requires a better hosting solution, which is not free, but as Ollama grows, this issue will become more and more important. Ollama teams could look for a sponsor who would pay for the hosting.

<!-- gh-comment-id:1816028235 --> @igorschlum commented on GitHub (Nov 17, 2023): You raise a good point, @priamai. Having a better bandwidth requires a better hosting solution, which is not free, but as Ollama grows, this issue will become more and more important. Ollama teams could look for a sponsor who would pay for the hosting.
GiteaMirror commented 2026-04-12 10:17:26 -05:00
Author
Owner
Copy Link

@orlyandico commented on GitHub (Nov 17, 2023):

I also tested this on AWS and Hyperstack instances, and didn't get the EOF retries. So it appears to be down to one's internet connection.

My more pressing question is why does Ollama have its own model repository, rather than supporting downloading models from, say, HuggingFace.

<!-- gh-comment-id:1816121142 --> @orlyandico commented on GitHub (Nov 17, 2023): I also tested this on AWS and Hyperstack instances, and didn't get the EOF retries. So it appears to be down to one's internet connection. My more pressing question is why does Ollama have its own model repository, rather than supporting downloading models from, say, HuggingFace.
GiteaMirror commented 2026-04-12 10:17:27 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Nov 17, 2023):

Hi folks thanks for creating an issue – looking into this – there really shouldn't be any EOF errors

<!-- gh-comment-id:1816311884 --> @jmorganca commented on GitHub (Nov 17, 2023): Hi folks thanks for creating an issue – looking into this – there really shouldn't be any EOF errors
GiteaMirror commented 2026-04-12 10:17:28 -05:00
Author
Owner
Copy Link

@priamai commented on GitHub (Nov 17, 2023):

We have some spare capacity on Aws S3 and also happy to sponsor. Send me a PM if you like. cheers.

<!-- gh-comment-id:1816420074 --> @priamai commented on GitHub (Nov 17, 2023): We have some spare capacity on Aws S3 and also happy to sponsor. Send me a PM if you like. cheers.
GiteaMirror commented 2026-04-12 10:17:29 -05:00
Author
Owner
Copy Link

@BruceMacD commented on GitHub (Mar 11, 2024):

Thanks for the report here. The bug shown in this issue specifically is now fixed as we run models directly rather than in a subprocess. Although there are still on EOF errors. If anyone else sees an EOF please open a new issue so we can triage it appropriately.

<!-- gh-comment-id:1989152214 --> @BruceMacD commented on GitHub (Mar 11, 2024): Thanks for the report here. The bug shown in this issue specifically is now fixed as we run models directly rather than in a subprocess. Although there are still on EOF errors. If anyone else sees an EOF please open a new issue so we can triage it appropriately.
GiteaMirror commented 2026-04-12 10:17:29 -05:00
Author
Owner
Copy Link

@MADAO-LUV commented on GitHub (Nov 12, 2024):

i also have this the same issue.And how i can solve this problem.i try reload for three times.

<!-- gh-comment-id:2469587643 --> @MADAO-LUV commented on GitHub (Nov 12, 2024): i also have this the same issue.And how i can solve this problem.i try reload for three times.
GiteaMirror commented 2026-04-12 10:17:30 -05:00
Author
Owner
Copy Link

@MichelMichels commented on GitHub (Jan 28, 2025):

Just reporting that I had this issue multiple times while trying to download deepseek-r1:7b.

EDIT: Currently stuck at 4.1 GB/4.7 GB

<!-- gh-comment-id:2618769939 --> @MichelMichels commented on GitHub (Jan 28, 2025): Just reporting that I had this issue multiple times while trying to download deepseek-r1:7b. EDIT: Currently stuck at 4.1 GB/4.7 GB
GiteaMirror commented 2026-04-12 10:17:30 -05:00
Author
Owner
Copy Link

@LeGao-HIT commented on GitHub (Jan 28, 2025):

Just reporting that I had this issue multiple times while trying to download deepseek-r1:7b

Agree, I also encountered the same problem when downloading deepseek-r1:7b: Error: max retries exceeded: EOF

<!-- gh-comment-id:2618775772 --> @LeGao-HIT commented on GitHub (Jan 28, 2025): > Just reporting that I had this issue multiple times while trying to download deepseek-r1:7b Agree, I also encountered the same problem when downloading deepseek-r1:7b: Error: max retries exceeded: EOF
GiteaMirror commented 2026-04-12 10:17:31 -05:00
Author
Owner
Copy Link

@muhamedljatifi commented on GitHub (Jan 28, 2025):

yes have the same issue with the same model. other models work fine

<!-- gh-comment-id:2618830947 --> @muhamedljatifi commented on GitHub (Jan 28, 2025): yes have the same issue with the same model. other models work fine
GiteaMirror commented 2026-04-12 10:17:32 -05:00
Author
Owner
Copy Link

@MengyangGao commented on GitHub (Jan 28, 2025):

same problem when ollama run deepseek-r1 which is from https://ollama.com/library/deepseek-r1

<!-- gh-comment-id:2619007726 --> @MengyangGao commented on GitHub (Jan 28, 2025): same problem when `ollama run deepseek-r1` which is from https://ollama.com/library/deepseek-r1
GiteaMirror commented 2026-04-12 10:17:32 -05:00
Author
Owner
Copy Link

@youkuzazazazaza commented on GitHub (Jan 28, 2025):

是的，同一型号有同样的问题。其他型号运行良好

yes have the same issue with the same model. other models work fine

i change to deepseek-r1:8b,it works well

<!-- gh-comment-id:2619229984 --> @youkuzazazazaza commented on GitHub (Jan 28, 2025): > 是的，同一型号有同样的问题。其他型号运行良好 > yes have the same issue with the same model. other models work fine i change to deepseek-r1:8b,it works well
GiteaMirror commented 2026-04-12 10:17:33 -05:00
Author
Owner
Copy Link

@micokonsep commented on GitHub (Jan 28, 2025):

try this.

https://medium.com/@micowendy/ollama-error-max-retries-exceeded-eof-c8baf4a57ffc
it;s in Indonesian language, but you can copy paste the code.

regards,

<!-- gh-comment-id:2619445783 --> @micokonsep commented on GitHub (Jan 28, 2025): try this. https://medium.com/@micowendy/ollama-error-max-retries-exceeded-eof-c8baf4a57ffc it;s in Indonesian language, but you can copy paste the code. regards,
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#587
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?