github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-07-16 21:25:46 -05:00
Code Issues 16.9k Packages Projects Releases 122 Wiki Activity

[GH-ISSUE #2111] Enable installation without root priviledge #26969

New Issue
Closed
opened 2026-04-22 03:46:42 -05:00 by GiteaMirror · 15 comments
GiteaMirror commented 2026-04-22 03:46:42 -05:00
Owner
Copy Link

Originally created by @chunhualiao on GitHub (Jan 20, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/2111

It seems like ollama will run sudo during its installation on Linux.

Please support the installation and use of users without sudo priviledge.

thanks.

Originally created by @chunhualiao on GitHub (Jan 20, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/2111 It seems like ollama will run sudo during its installation on Linux. Please support the installation and use of users without sudo priviledge. thanks.
GiteaMirror commented 2026-04-22 03:46:43 -05:00
Author
Owner
Copy Link

@jmorganca commented on GitHub (Jan 21, 2024):

Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an ollama user. The releases include pre-built binaries that will work without any sudo (e.g. just run ollama serve). Let me know if that helps!

<!-- gh-comment-id:1902449564 --> @jmorganca commented on GitHub (Jan 21, 2024): Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an `ollama` user. The [releases](https://github.com/jmorganca/ollama/releases) include pre-built binaries that will work without any sudo (e.g. just run `ollama serve`). Let me know if that helps!
GiteaMirror commented 2026-04-22 03:46:43 -05:00
Author
Owner
Copy Link

@ReanFernandes commented on GitHub (Jan 21, 2024):

Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an ollama user. The releases include pre-built binaries that will work without any sudo (e.g. just run ollama serve). Let me know if that helps!

Hello!
Im trying to get ollama up and running on a cluster which i do not have sudo access to. Could you please elaborate how I could go about the installation?
Note: I am a newbie to this, and the install page does not have any info on how to go about this. I would gladly appreciate any help you could provide! :)
Thank you!

<!-- gh-comment-id:1902751675 --> @ReanFernandes commented on GitHub (Jan 21, 2024): > Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an `ollama` user. The [releases](https://github.com/jmorganca/ollama/releases) include pre-built binaries that will work without any sudo (e.g. just run `ollama serve`). Let me know if that helps! Hello! Im trying to get ollama up and running on a cluster which i do not have sudo access to. Could you please elaborate how I could go about the installation? Note: I am a newbie to this, and the [install page](https://github.com/jmorganca/ollama/blob/main/docs/linux.md) does not have any info on how to go about this. I would gladly appreciate any help you could provide! :) Thank you!
GiteaMirror commented 2026-04-22 03:46:44 -05:00
Author
Owner
Copy Link

@chunhualiao commented on GitHub (Feb 9, 2024):

@ReanFernandes The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64.

Just download it to your Linux cluster, then run the following:

start the server in background

./ollama-linux-amd64 serve&

run a local model afterwards

./ollama-linux-amd64 run llama2

I wish someone can add this into their official instructions.

<!-- gh-comment-id:1935146890 --> @chunhualiao commented on GitHub (Feb 9, 2024): @ReanFernandes The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64. Just download it to your Linux cluster, then run the following: # start the server in background ./ollama-linux-amd64 serve& # run a local model afterwards ./ollama-linux-amd64 run llama2 I wish someone can add this into their official instructions.
GiteaMirror commented 2026-04-22 03:46:45 -05:00
Author
Owner
Copy Link

@Anirudh257 commented on GitHub (Feb 21, 2024):

@ReanFernandes The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64.

Just download it to your Linux cluster, then run the following:

start the server in background

./ollama-linux-amd64 serve&

run a local model afterwards

./ollama-linux-amd64 run llama2

I wish someone can add this into their official instructions.

Hi @chunhualiao even bash access is prohibited on cluster. @ReanFernandes you will need to build it from scratch on your cluster.

<!-- gh-comment-id:1955463190 --> @Anirudh257 commented on GitHub (Feb 21, 2024): > @ReanFernandes The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64. > > Just download it to your Linux cluster, then run the following: > > # start the server in background > ./ollama-linux-amd64 serve& > > # run a local model afterwards > ./ollama-linux-amd64 run llama2 > > I wish someone can add this into their official instructions. Hi @chunhualiao even bash access is prohibited on cluster. @ReanFernandes you will need to build it from scratch on your cluster.
GiteaMirror commented 2026-04-22 03:46:46 -05:00
Author
Owner
Copy Link

@zrgtzjwh commented on GitHub (Mar 14, 2024):

@ReanFernandes The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64.

Just download it to your Linux cluster, then run the following:

start the server in background

./ollama-linux-amd64 serve&

run a local model afterwards

./ollama-linux-amd64 run llama2

I wish someone can add this into their official instructions.

It works for me.

<!-- gh-comment-id:1996640855 --> @zrgtzjwh commented on GitHub (Mar 14, 2024): > @ReanFernandes The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64. > > Just download it to your Linux cluster, then run the following: > > # start the server in background > ./ollama-linux-amd64 serve& > > # run a local model afterwards > ./ollama-linux-amd64 run llama2 > > I wish someone can add this into their official instructions. It works for me.
GiteaMirror commented 2026-04-22 03:46:49 -05:00
Author
Owner
Copy Link

@FinlandBreakfast commented on GitHub (Mar 28, 2024):

The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64.

Just download it to your Linux cluster, then run the following:

start the server in background

./ollama-linux-amd64 serve&

run a local model afterwards

./ollama-linux-amd64 run llama2

I wish someone can add this into their official instructions.

Build from scratch instructions?

<!-- gh-comment-id:2024204028 --> @FinlandBreakfast commented on GitHub (Mar 28, 2024): > The download page has a list of assets, one of them is binary for Linux named ollama-linux-amd64. > > Just download it to your Linux cluster, then run the following: > > # start the server in background > ./ollama-linux-amd64 serve& > > # run a local model afterwards > ./ollama-linux-amd64 run llama2 > > I wish someone can add this into their official instructions. Build from scratch instructions?
GiteaMirror commented 2026-04-22 03:46:51 -05:00
Author
Owner
Copy Link

@bozo32 commented on GitHub (Apr 5, 2024):

had to
chmod +x ollama-*
to get it to work

<!-- gh-comment-id:2039783492 --> @bozo32 commented on GitHub (Apr 5, 2024): had to chmod +x ollama-* to get it to work
GiteaMirror commented 2026-04-22 03:46:52 -05:00
Author
Owner
Copy Link

@krz commented on GitHub (Apr 26, 2024):

I'm in a similar situation, trying to run ollama without sudo access. Downloading ollama and starting the server with ./ollama-linux-amd64 serve& worked fine, but when I try to run a model, I get this error:

time=2024-04-26T15:13:33.513+02:00 level=INFO source=gpu.go:126 msg="Nvidia GPU detected via cudart"
time=2024-04-26T15:13:33.513+02:00 level=INFO source=cpu_common.go:11 msg="CPU has AVX2"
⠙ time=2024-04-26T15:13:33.588+02:00 level=INFO source=gpu.go:202 msg="[cudart] CUDART CUDA Compute Capability detected: 6.0"
time=2024-04-26T15:13:33.618+02:00 level=INFO source=server.go:127 msg="offload to gpu" reallayers=33 layers=33 required="5222.5 MiB" used="5222.5 MiB" available="16020.1 MiB" kv="1024.0 MiB" fulloffload="164.0 MiB" partialoffload="193.0 MiB"
time=2024-04-26T15:13:33.618+02:00 level=INFO source=cpu_common.go:11 msg="CPU has AVX2"
time=2024-04-26T15:13:33.622+02:00 level=INFO source=server.go:264 msg="starting llama server" cmd="/tmp/ollama2278607108/runners/cuda_v11/ollama_llama_server --model /home/oll/.ollama/models/blobs/sha256-8934d96d3f08982e95922b2b7a2c626a1fe873d7c3b06e8e56d7bc0a1fef9246 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 33 --port 45845"
time=2024-04-26T15:13:33.623+02:00 level=INFO source=server.go:264 msg="starting llama server" cmd="/tmp/ollama2278607108/runners/cpu_avx2/ollama_llama_server --model /home/oll/.ollama/models/blobs/sha256-8934d96d3f08982e95922b2b7a2c626a1fe873d7c3b06e8e56d7bc0a1fef9246 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 33 --port 40273"
time=2024-04-26T15:13:33.624+02:00 level=ERROR source=server.go:285 msg="unable to load any llama server" error="error starting the external llama server: fork/exec /tmp/ollama2278607108/runners/cpu_avx2/ollama_llama_server: permission denied "
[GIN] 2024/04/26 - 15:13:33 | 500 | 1.176409787s | 127.0.0.1 | POST "/api/chat"
Error: error starting the external llama server: fork/exec /tmp/ollama2278607108/runners/cpu_avx2/ollama_llama_server: permission denied

<!-- gh-comment-id:2079380112 --> @krz commented on GitHub (Apr 26, 2024): I'm in a similar situation, trying to run ollama without sudo access. Downloading ollama and starting the server with `./ollama-linux-amd64 serve&` worked fine, but when I try to run a model, I get this error: > time=2024-04-26T15:13:33.513+02:00 level=INFO source=gpu.go:126 msg="Nvidia GPU detected via cudart" time=2024-04-26T15:13:33.513+02:00 level=INFO source=cpu_common.go:11 msg="CPU has AVX2" ⠙ time=2024-04-26T15:13:33.588+02:00 level=INFO source=gpu.go:202 msg="[cudart] CUDART CUDA Compute Capability detected: 6.0" time=2024-04-26T15:13:33.618+02:00 level=INFO source=server.go:127 msg="offload to gpu" reallayers=33 layers=33 required="5222.5 MiB" used="5222.5 MiB" available="16020.1 MiB" kv="1024.0 MiB" fulloffload="164.0 MiB" partialoffload="193.0 MiB" time=2024-04-26T15:13:33.618+02:00 level=INFO source=cpu_common.go:11 msg="CPU has AVX2" time=2024-04-26T15:13:33.622+02:00 level=INFO source=server.go:264 msg="starting llama server" cmd="/tmp/ollama2278607108/runners/cuda_v11/ollama_llama_server --model /home/oll/.ollama/models/blobs/sha256-8934d96d3f08982e95922b2b7a2c626a1fe873d7c3b06e8e56d7bc0a1fef9246 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 33 --port 45845" time=2024-04-26T15:13:33.623+02:00 level=INFO source=server.go:264 msg="starting llama server" cmd="/tmp/ollama2278607108/runners/cpu_avx2/ollama_llama_server --model /home/oll/.ollama/models/blobs/sha256-8934d96d3f08982e95922b2b7a2c626a1fe873d7c3b06e8e56d7bc0a1fef9246 --ctx-size 2048 --batch-size 512 --embedding --log-disable --n-gpu-layers 33 --port 40273" time=2024-04-26T15:13:33.624+02:00 level=ERROR source=server.go:285 msg="unable to load any llama server" error="error starting the external llama server: fork/exec /tmp/ollama2278607108/runners/cpu_avx2/ollama_llama_server: permission denied " [GIN] 2024/04/26 - 15:13:33 | 500 | 1.176409787s | 127.0.0.1 | POST "/api/chat" Error: error starting the external llama server: fork/exec /tmp/ollama2278607108/runners/cpu_avx2/ollama_llama_server: permission denied
GiteaMirror commented 2026-04-22 03:46:54 -05:00
Author
Owner
Copy Link

@haozhuang0000 commented on GitHub (Jul 2, 2024):

I got this error by following above comments, anyone know how?

Error: pull model manifest: Get "https://registry.ollama.ai/v2/library/llama3/manifests/8b": read tcp xxxxxx: read: connection reset by peer

<!-- gh-comment-id:2202550650 --> @haozhuang0000 commented on GitHub (Jul 2, 2024): I got this error by following above comments, anyone know how? Error: pull model manifest: Get "https://registry.ollama.ai/v2/library/llama3/manifests/8b": read tcp xxxxxx: read: connection reset by peer
GiteaMirror commented 2026-04-22 03:46:55 -05:00
Author
Owner
Copy Link

@GonyRosenman commented on GitHub (Jul 15, 2024):

i used chatgpt to modify the install script (can anyone verify this solution???) -
#!/bin/sh
set -eu

status() { echo ">>> $" >&2; }
error() { echo "ERROR $"; exit 1; }
warning() { echo "WARNING: $*"; }

TEMP_DIR=$(mktemp -d)
cleanup() { rm -rf $TEMP_DIR; }
trap cleanup EXIT

ARCH=$(uname -m)
case "$ARCH" in
x86_64) ARCH="amd64" ;;
aarch64|arm64) ARCH="arm64" ;;
*) error "Unsupported architecture: $ARCH" ;;
esac

VER_PARAM="${OLLAMA_VERSION:+?version=$OLLAMA_VERSION}"

status "Downloading ollama..."
curl --fail --show-error --location --progress-bar -o $TEMP_DIR/ollama "https://ollama.com/download/ollama-linux-${ARCH}${VER_PARAM}"

Ensure bin directory exists

mkdir -p ~/bin
status "Installing ollama to ~/bin..."
install -m755 $TEMP_DIR/ollama ~/bin/ollama

status 'Install complete. Run "~/bin/ollama" from the command line.'

<!-- gh-comment-id:2228480051 --> @GonyRosenman commented on GitHub (Jul 15, 2024): i used chatgpt to modify the install script (can anyone verify this solution???) - #!/bin/sh set -eu status() { echo ">>> $*" >&2; } error() { echo "ERROR $*"; exit 1; } warning() { echo "WARNING: $*"; } TEMP_DIR=$(mktemp -d) cleanup() { rm -rf $TEMP_DIR; } trap cleanup EXIT ARCH=$(uname -m) case "$ARCH" in x86_64) ARCH="amd64" ;; aarch64|arm64) ARCH="arm64" ;; *) error "Unsupported architecture: $ARCH" ;; esac VER_PARAM="${OLLAMA_VERSION:+?version=$OLLAMA_VERSION}" status "Downloading ollama..." curl --fail --show-error --location --progress-bar -o $TEMP_DIR/ollama "https://ollama.com/download/ollama-linux-${ARCH}${VER_PARAM}" # Ensure bin directory exists mkdir -p ~/bin status "Installing ollama to ~/bin..." install -m755 $TEMP_DIR/ollama ~/bin/ollama status 'Install complete. Run "~/bin/ollama" from the command line.'
GiteaMirror commented 2026-04-22 03:46:56 -05:00
Author
Owner
Copy Link

@justin4ai commented on GitHub (Oct 3, 2024):

@chunhualiao Hello, thanks for your information!

But running your commands actually doesn't work for me - I tried it after I unzip .tgz:

./ollama-linux-amd64 serve&
[1] 15739
bash: ./ollama-linux-amd64: No such file or directory
[1]+  Exit 127

My folder tree looks like,
image
.

Do you have any clue? If so, I'll be so glad if you let me know!

<!-- gh-comment-id:2391778255 --> @justin4ai commented on GitHub (Oct 3, 2024): @chunhualiao Hello, thanks for your information! But running your commands actually doesn't work for me - I tried it after I unzip .tgz: ``` ./ollama-linux-amd64 serve& [1] 15739 bash: ./ollama-linux-amd64: No such file or directory [1]+ Exit 127 ``` My folder tree looks like, ![image](https://github.com/user-attachments/assets/28522aa3-3d82-46fc-976e-9921c7ea9288) . Do you have any clue? If so, I'll be so glad if you let me know!
GiteaMirror commented 2026-04-22 03:46:56 -05:00
Author
Owner
Copy Link

@Haruki-Kny commented on GitHub (Oct 8, 2024):

@justin4ai
Hi, I also had the same problem.
I think you should execute ./bin/ollama serve&. It's because the name of the extracting directory might have changed from ollama-linux-amd64 to bin.
If it doesn't work, you should check the content of the bin.

<!-- gh-comment-id:2399617839 --> @Haruki-Kny commented on GitHub (Oct 8, 2024): @justin4ai Hi, I also had the same problem. I think you should execute `./bin/ollama serve&`. It's because the name of the extracting directory might have changed from `ollama-linux-amd64` to `bin`. If it doesn't work, you should check the content of the bin.
GiteaMirror commented 2026-04-22 03:46:57 -05:00
Author
Owner
Copy Link

@famura commented on GitHub (Feb 20, 2025):

Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an ollama user. The releases include pre-built binaries that will work without any sudo (e.g. just run ollama serve). Let me know if that helps!

Can you please say more about this? On this doc page I still read sudo before tar.

Let's assume I am under Linux, have no sudo permissions, and I just downloaded v0.5.10.tar.gz. How do I get to run ollama such that I can use it in a different process, e.g. from a python script via the ollama python package?

<!-- gh-comment-id:2672183889 --> @famura commented on GitHub (Feb 20, 2025): > Hi [@chunhualiao](https://github.com/chunhualiao), thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an `ollama` user. The [releases](https://github.com/jmorganca/ollama/releases) include pre-built binaries that will work without any sudo (e.g. just run `ollama serve`). Let me know if that helps! Can you please say more about this? On [this doc page](https://github.com/ollama/ollama/blob/main/docs/linux.md) I still read `sudo` before tar. Let's assume I am under Linux, have no sudo permissions, and I just downloaded v0.5.10.tar.gz. How do I get to run ollama such that I can use it in a different process, e.g. from a python script via the ollama python package?
GiteaMirror commented 2026-04-22 03:46:57 -05:00
Author
Owner
Copy Link

@wangjiawen2013 commented on GitHub (Mar 27, 2025):

Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an ollama user. The releases include pre-built binaries that will work without any sudo (e.g. just run ollama serve). Let me know if that helps!

Is there a binary for Centos linux server ? I only find binary for macos.

<!-- gh-comment-id:2756488366 --> @wangjiawen2013 commented on GitHub (Mar 27, 2025): > Hi [@chunhualiao](https://github.com/chunhualiao), thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an `ollama` user. The [releases](https://github.com/jmorganca/ollama/releases) include pre-built binaries that will work without any sudo (e.g. just run `ollama serve`). Let me know if that helps! Is there a binary for Centos linux server ? I only find binary for macos.
GiteaMirror commented 2026-04-22 03:46:58 -05:00
Author
Owner
Copy Link

@mobraine commented on GitHub (Jul 19, 2025):

Hi @chunhualiao, thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an ollama user. The releases include pre-built binaries that will work without any sudo (e.g. just run ollama serve). Let me know if that helps!

Can you please say more about this? On this doc page I still read sudo before tar.

Let's assume I am under Linux, have no sudo permissions, and I just downloaded v0.5.10.tar.gz. How do I get to run ollama such that I can use it in a different process, e.g. from a python script via the ollama python package?

If you have downloaded the tgz file, simply tar -xzvf v0.5.10.tar.gz this should extract and show you where the files are extracted. Like @justin4ai said, ollama is now living in the bin directory.

@justin4ai Hi, I also had the same problem. I think you should execute ./bin/ollama serve&. It's because the name of the extracting directory might have changed from ollama-linux-amd64 to bin. If it doesn't work, you should check the content of the bin.

<!-- gh-comment-id:3092620891 --> @mobraine commented on GitHub (Jul 19, 2025): > > Hi [@chunhualiao](https://github.com/chunhualiao), thanks for the issue. The install script does this because it needs to check for devices, drivers and set up an `ollama` user. The [releases](https://github.com/jmorganca/ollama/releases) include pre-built binaries that will work without any sudo (e.g. just run `ollama serve`). Let me know if that helps! > > Can you please say more about this? On [this doc page](https://github.com/ollama/ollama/blob/main/docs/linux.md) I still read `sudo` before tar. > > Let's assume I am under Linux, have no sudo permissions, and I just downloaded v0.5.10.tar.gz. How do I get to run ollama such that I can use it in a different process, e.g. from a python script via the ollama python package? If you have downloaded the tgz file, simply `tar -xzvf v0.5.10.tar.gz` this should extract and show you where the files are extracted. Like @justin4ai said, ollama is now living in the bin directory. > [@justin4ai](https://github.com/justin4ai) Hi, I also had the same problem. I think you should execute `./bin/ollama serve&`. It's because the name of the extracting directory might have changed from `ollama-linux-amd64` to `bin`. If it doesn't work, you should check the content of the bin.
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
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#26969
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?