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 #1938] ollama --version 0.1.20 not working #63152

New Issue
Closed
opened 2026-05-03 12:18:03 -05:00 by GiteaMirror · 14 comments
GiteaMirror commented 2026-05-03 12:18:03 -05:00
Owner
Copy Link

Originally created by @PhilipAmadasun on GitHub (Jan 11, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/1938

Originally assigned to: @jmorganca on GitHub.

Our ollama no longer works once upgrading to version 0.1.20. All the commands, for instance:

curl http://localhost:11434/api/chat -d '{
>   "model": "llama2",
>   "messages": [
>     {
>       "role": "user",
>       "content": "why is the sky blue?"
>     }
>   ]
> }'

Just gets stuck and doesn't run. What's going on? I believe this is the latest version, is the version not stable? Do we have to downgrade ollama? If so how do we go about doing that?

Originally created by @PhilipAmadasun on GitHub (Jan 11, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/1938 Originally assigned to: @jmorganca on GitHub. Our ollama no longer works once upgrading to version `0.1.20`. All the commands, for instance: ``` curl http://localhost:11434/api/chat -d '{ > "model": "llama2", > "messages": [ > { > "role": "user", > "content": "why is the sky blue?" > } > ] > }' ``` Just gets stuck and doesn't run. What's going on? I believe this is the latest version, is the version not stable? Do we have to downgrade ollama? If so how do we go about doing that?
GiteaMirror added the bug label 2026-05-03 12:18:03 -05:00
GiteaMirror commented 2026-05-03 12:18:06 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Jan 12, 2024):

Hi @PhilipAmadasun - I'm sorry it's hanging for you. You definitely shouldn't need to downgrade – 0.1.20 was focused on stability around CUDA, although there's still a bit more work to on it. To help me track it down:

  • Is this on macOS or Linux?
  • If Linux, what kind of GPU?
  • Do you have the logs handy? journalctl --no-pager -u ollama on Linux and cat ~/.ollama/logs/server.log on macOS

Thanks so much

<!-- gh-comment-id:1888262045 --> @jmorganca commented on GitHub (Jan 12, 2024): Hi @PhilipAmadasun - I'm sorry it's hanging for you. You definitely shouldn't need to downgrade – `0.1.20` was focused on stability around CUDA, although there's still a bit more work to on it. To help me track it down: - Is this on macOS or Linux? - If Linux, what kind of GPU? - Do you have the logs handy? `journalctl --no-pager -u ollama` on Linux and `cat ~/.ollama/logs/server.log` on macOS Thanks so much
GiteaMirror commented 2026-05-03 12:18:06 -05:00
Author
Owner
Copy Link

@PhilipAmadasun commented on GitHub (Jan 12, 2024):

Here are the logs:
ollama_logs.txt

<!-- gh-comment-id:1889825695 --> @PhilipAmadasun commented on GitHub (Jan 12, 2024): Here are the logs: [ollama_logs.txt](https://github.com/jmorganca/ollama/files/13923196/ollama_logs.txt)
GiteaMirror commented 2026-05-03 12:18:07 -05:00
Author
Owner
Copy Link

@PhilipAmadasun commented on GitHub (Jan 12, 2024):

@jmorganca We're using linux (Ubuntu 22.04). These are the GPU specs ```
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.147.05 Driver Version: 525.147.05 CUDA Version: 12.0 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 Tesla P100-PCIE... On | 00000000:03:00.0 Off | 0 |
| N/A 30C P0 27W / 250W | 0MiB / 16384MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
| 1 Tesla P100-PCIE... On | 00000000:82:00.0 Off | 0 |
| N/A 32C P0 26W / 250W | 0MiB / 16384MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+

<!-- gh-comment-id:1890048470 --> @PhilipAmadasun commented on GitHub (Jan 12, 2024): @jmorganca We're using linux (Ubuntu 22.04). These are the GPU specs ``` +-----------------------------------------------------------------------------+ | NVIDIA-SMI 525.147.05 Driver Version: 525.147.05 CUDA Version: 12.0 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |===============================+======================+======================| | 0 Tesla P100-PCIE... On | 00000000:03:00.0 Off | 0 | | N/A 30C P0 27W / 250W | 0MiB / 16384MiB | 0% Default | | | | N/A | +-------------------------------+----------------------+----------------------+ | 1 Tesla P100-PCIE... On | 00000000:82:00.0 Off | 0 | | N/A 32C P0 26W / 250W | 0MiB / 16384MiB | 0% Default | | | | N/A | +-------------------------------+----------------------+----------------------+ ```
GiteaMirror commented 2026-05-03 12:18:08 -05:00
Author
Owner
Copy Link

@Rust-Ninja-Sabi commented on GitHub (Jan 14, 2024):

I think it the same bug:

>ollama run mixtral     
zsh: illegal hardware instruction  ollama run mixtral
<!-- gh-comment-id:1891011275 --> @Rust-Ninja-Sabi commented on GitHub (Jan 14, 2024): I think it the same bug: ``` >ollama run mixtral zsh: illegal hardware instruction ollama run mixtral ```
GiteaMirror commented 2026-05-03 12:18:09 -05:00
Author
Owner
Copy Link

@Rust-Ninja-Sabi commented on GitHub (Jan 14, 2024):

Is it possible to download a older version?

<!-- gh-comment-id:1891011598 --> @Rust-Ninja-Sabi commented on GitHub (Jan 14, 2024): Is it possible to download a older version?
GiteaMirror commented 2026-05-03 12:18:10 -05:00
Author
Owner
Copy Link

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

@Rust-Ninja-Sabi Yes use command:

curl https://ollama.ai/install.sh | sed 's#https://ollama.ai/download#https://github.com/jmorganca/ollama/releases/download/v0.1.17#' | sh
<!-- gh-comment-id:1891037817 --> @PhilipAmadasun commented on GitHub (Jan 14, 2024): @Rust-Ninja-Sabi Yes use command: ``` curl https://ollama.ai/install.sh | sed 's#https://ollama.ai/download#https://github.com/jmorganca/ollama/releases/download/v0.1.17#' | sh ```
GiteaMirror commented 2026-05-03 12:18:11 -05:00
Author
Owner
Copy Link

@Rust-Ninja-Sabi commented on GitHub (Jan 17, 2024):

Thanks. This script does not run on macOS

<!-- gh-comment-id:1895317633 --> @Rust-Ninja-Sabi commented on GitHub (Jan 17, 2024): Thanks. This script does not run on macOS
GiteaMirror commented 2026-05-03 12:18:12 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jan 26, 2024):

Here's the excerpt from the log where it wen't bad.

Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: llama_apply_lora_from_file_internal: applying lora adapter from '/usr/share/ollama/.ollama/models/blobs/sha256:f4e82fc0919ab5e92b0bf8230154a96cd6c0462a7583b39af0ab6f4d1c8d3521' - please wait ...
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: llama_apply_lora_from_file_internal: bad file magic
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: llama_init_from_gpt_params: error: failed to apply lora adapter
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: Lazy loading /tmp/ollama2924267924/cuda/libext_server.so library
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: Lazy loading /tmp/ollama2924267924/cuda/libext_server.so library
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: {"timestamp":1705015928,"level":"ERROR","function":"load_model","line":581,"message":"unable to load model","model":"/usr/share/ollama/.ollama/models/blobs/sha256:e8a35b5937a5e6d5c35d1f2a15f161e07eefe5e5bb0a3cdd42998ee79b057730"}
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: 2024/01/11 23:32:08 llm.go:129: Failed to load dynamic library cuda - falling back to CPU mode error loading model /usr/share/ollama/.ollama/models/blobs/sha256:e8a35b5937a5e6d5c35d1f2a15f161e07eefe5e5bb0a3cdd42998ee79b057
Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: 2024/01/11 23:32:08 ext_server_common.go:85: concurrent llm servers not yet supported, waiting for prior server to complete
Jan 12 18:53:33 arnold.ailab.internal systemd[1]: Stopping Ollama Service...

I can't speak to the lora adapter load problem, but that failure cascaded to another bug where we didn't unlock a lock and that lead to concurrent llm servers not yet supported, waiting for prior server to complete which was fixed a week ago. Upgrading to 0.1.22 will resolve the lock bug, but you might want to re-pull your models in case something got corrupted on your filesystem.

<!-- gh-comment-id:1912752547 --> @dhiltgen commented on GitHub (Jan 26, 2024): Here's the excerpt from the log where it wen't bad. ``` Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: llama_apply_lora_from_file_internal: applying lora adapter from '/usr/share/ollama/.ollama/models/blobs/sha256:f4e82fc0919ab5e92b0bf8230154a96cd6c0462a7583b39af0ab6f4d1c8d3521' - please wait ... Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: llama_apply_lora_from_file_internal: bad file magic Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: llama_init_from_gpt_params: error: failed to apply lora adapter Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: Lazy loading /tmp/ollama2924267924/cuda/libext_server.so library Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: Lazy loading /tmp/ollama2924267924/cuda/libext_server.so library Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: {"timestamp":1705015928,"level":"ERROR","function":"load_model","line":581,"message":"unable to load model","model":"/usr/share/ollama/.ollama/models/blobs/sha256:e8a35b5937a5e6d5c35d1f2a15f161e07eefe5e5bb0a3cdd42998ee79b057730"} Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: 2024/01/11 23:32:08 llm.go:129: Failed to load dynamic library cuda - falling back to CPU mode error loading model /usr/share/ollama/.ollama/models/blobs/sha256:e8a35b5937a5e6d5c35d1f2a15f161e07eefe5e5bb0a3cdd42998ee79b057 Jan 11 23:32:08 arnold.ailab.internal ollama[340593]: 2024/01/11 23:32:08 ext_server_common.go:85: concurrent llm servers not yet supported, waiting for prior server to complete Jan 12 18:53:33 arnold.ailab.internal systemd[1]: Stopping Ollama Service... ``` I can't speak to the lora adapter load problem, but that failure cascaded to another bug where we didn't unlock a lock and that lead to `concurrent llm servers not yet supported, waiting for prior server to complete` which was fixed a week ago. Upgrading to 0.1.22 will resolve the lock bug, but you might want to re-pull your models in case something got corrupted on your filesystem.
GiteaMirror commented 2026-05-03 12:18:12 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jan 26, 2024):

@Rust-Ninja-Sabi your problem is unrelated to this issue. You are most likely trying to run Ollama under Rosetta on an ARM mac, which until recently wasn't supported (resulting in an "illegal instruction" error). If you ugprade, it will work, but you should run Ollama as a native ARM app and you'll get much better performance.

<!-- gh-comment-id:1912754556 --> @dhiltgen commented on GitHub (Jan 26, 2024): @Rust-Ninja-Sabi your problem is unrelated to this issue. You are most likely trying to run Ollama under Rosetta on an ARM mac, which until recently wasn't supported (resulting in an "illegal instruction" error). If you ugprade, it will work, but you should run Ollama as a native ARM app and you'll get much better performance.
GiteaMirror commented 2026-05-03 12:18:13 -05:00
Author
Owner
Copy Link

@Rust-Ninja-Sabi commented on GitHub (Jan 27, 2024):

@dhiltgen hello Daniel
Thanks for your message. I installed Ollama again (version 0.1.22). Now it works. I installed it from Ollama homepage. I hope it is the native version.

<!-- gh-comment-id:1913093759 --> @Rust-Ninja-Sabi commented on GitHub (Jan 27, 2024): @dhiltgen hello Daniel Thanks for your message. I installed Ollama again (version 0.1.22). Now it works. I installed it from Ollama homepage. I hope it is the native version.
GiteaMirror commented 2026-05-03 12:18:14 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jan 27, 2024):

@Rust-Ninja-Sabi we compile as a "Mach-O universal binary" so a single executable contains both x86 and ARM variants and MacOS will pick the right one based on your configuration. Running under Rosetta will work now (where it used to crash), but will have a significant performance penalty.

<!-- gh-comment-id:1913216394 --> @dhiltgen commented on GitHub (Jan 27, 2024): @Rust-Ninja-Sabi we compile as a "Mach-O universal binary" so a single executable contains both x86 and ARM variants and MacOS will pick the right one based on your configuration. Running under Rosetta will work now (where it used to crash), but will have a significant performance penalty.
GiteaMirror commented 2026-05-03 12:18:14 -05:00
Author
Owner
Copy Link

@dhiltgen commented on GitHub (Jan 27, 2024):

@PhilipAmadasun please let us know if 0.1.22 resolves your problem

<!-- gh-comment-id:1913219546 --> @dhiltgen commented on GitHub (Jan 27, 2024): @PhilipAmadasun please let us know if 0.1.22 resolves your problem
GiteaMirror commented 2026-05-03 12:18:15 -05:00
Author
Owner
Copy Link

@Rust-Ninja-Sabi commented on GitHub (Jan 27, 2024):

Thanks. It is working.

<!-- gh-comment-id:1913249020 --> @Rust-Ninja-Sabi commented on GitHub (Jan 27, 2024): Thanks. It is working.
GiteaMirror commented 2026-05-03 12:18:16 -05:00
Author
Owner
Copy Link

@PhilipAmadasun commented on GitHub (Feb 16, 2024):

@dhiltgen @jmorganca All's good! Sorry or late response.

<!-- gh-comment-id:1947922011 --> @PhilipAmadasun commented on GitHub (Feb 16, 2024): @dhiltgen @jmorganca All's good! Sorry or late response.
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#63152
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?