0
0

国内网络在Ubuntu 22.04中在线安装Ollama并配置Open-Webui&Dify

配置docker科技网络

登录后复制

  1. 创建或编辑 Docker 配置文件 让docker使用代理：
  2. sudo mkdir /etc/systemd/system/docker.service.d -p
  3. sudo vim /etc/systemd/system/docker.service.d/http-proxy.conf 文件，并添加以下内容：
  4. [Service]
  5. Environment="HTTP_PROXY=http://10.10.9.232:30809"
  6. Environment="HTTPS_PROXY=http://10.10.9.232:30809"
  7. Environment="NO_PROXY=localhost,127.0.0.1"
  9. 重新加载 systemd 配置并重启 Docker 服务：
  10. sudo systemctl daemon-reload
  11. sudo systemctl restart docker
  13. 验证配置是否生效
  14. sudo systemctl show --property=Environment docker

Ubuntu 22.04 安装 Docker

登录后复制

  1. curl -fsSL https://get.docker.com | bash -s docker --mirror Aliyun
  2. systemctl enable --now docker

安装 docker-compose

登录后复制

  1. curl -L https://github.com/docker/compose/releases/download/v2.20.3/docker-compose-`uname -s`-`uname -m` -o /usr/local/bin/docker-compose
  2. chmod +x /usr/local/bin/docker-compose

验证安装

登录后复制

  1. docker -v
  2. docker-compose -v

Ubuntu 22.04 安装ollama

  • 在开始之前必须具备科技网络，然后安装并配置好proxychains
  • 配置好proxychains工具之后，将在线安装脚本的所有curl命令前面增加命令：proxychains
  • 然后正常执行安装脚本，期间保证科技网络能够使用，没有意外的话即可在线安装成功。

在线安装ollama

登录后复制

  1. 国内，安装因为速度太慢而超时安装不了，所以，将安装脚本下载下来，走代理安装
  2. wget https://ollama.com/install.sh
  3. chmod +x install.sh
  5. 或者
  6. curl -fsSL https://ollama.com/install.sh -o ollama_install.sh
  7. chmod +x ollama_install.sh
  9. 或者
  10. curl -O https://ollama.com/install.sh
  11. chmod +x install.sh

  • 修改版本原始安装脚本

登录后复制

  1. #!/bin/sh
  2. # This script installs Ollama on Linux.
  3. # It detects the current operating system architecture and installs the appropriate version of Ollama.
  5. set -eu
  7. status() { echo ">>> $*" >&2; }
  8. error() { echo "ERROR $*"; exit 1; }
  9. warning() { echo "WARNING: $*"; }
  11. TEMP_DIR=$(mktemp -d)
  12. cleanup() { rm -rf $TEMP_DIR; }
  13. trap cleanup EXIT
  15. available() { command -v $1 >/dev/null; }
  16. require() {
  17.     local MISSING=''
  18.     for TOOL in $*; do
  19.         if ! available $TOOL; then
  20.             MISSING="$MISSING $TOOL"
  21.         fi
  22.     done
  24.     echo $MISSING
  25. }
  27. [ "$(uname -s)" = "Linux" ] || error 'This script is intended to run on Linux only.'
  29. ARCH=$(uname -m)
  30. case "$ARCH" in
  31.     x86_64) ARCH="amd64" ;;
  32.     aarch64|arm64) ARCH="arm64" ;;
  33.     *) error "Unsupported architecture: $ARCH" ;;
  34. esac
  36. IS_WSL2=false
  38. KERN=$(uname -r)
  39. case "$KERN" in
  40.     *icrosoft*WSL2 | *icrosoft*wsl2) IS_WSL2=true;;
  41.     *icrosoft) error "Microsoft WSL1 is not currently supported. Please use WSL2 with 'wsl --set-version <distro> 2'" ;;
  42.     *) ;;
  43. esac
  45. VER_PARAM="${OLLAMA_VERSION:+?version=$OLLAMA_VERSION}"
  47. SUDO=
  48. if [ "$(id -u)" -ne 0 ]; then
  49.     # Running as root, no need for sudo
  50.     if ! available sudo; then
  51.         error "This script requires superuser permissions. Please re-run as root."
  52.     fi
  54.     SUDO="sudo"
  55. fi
  57. NEEDS=$(require curl awk grep sed tee xargs)
  58. if [ -n "$NEEDS" ]; then
  59.     status "ERROR: The following tools are required but missing:"
  60.     for NEED in $NEEDS; do
  61.         echo "  - $NEED"
  62.     done
  63.     exit 1
  64. fi
  66. for BINDIR in /usr/local/bin /usr/bin /bin; do
  67.     echo $PATH | grep -q $BINDIR && break || continue
  68. done
  69. OLLAMA_INSTALL_DIR=$(dirname ${BINDIR})
  71. status "Installing ollama to $OLLAMA_INSTALL_DIR"
  72. $SUDO install -o0 -g0 -m755 -d $BINDIR
  73. $SUDO install -o0 -g0 -m755 -d "$OLLAMA_INSTALL_DIR"
  74. if proxychains curl -I --silent --fail --location "https://ollama.com/download/ollama-linux-${ARCH}.tgz${VER_PARAM}" >/dev/null ; then
  75.     status "Downloading Linux ${ARCH} bundle"
  76.     proxychains curl --fail --show-error --location --progress-bar \
  77.         "https://ollama.com/download/ollama-linux-${ARCH}.tgz${VER_PARAM}" | \
  78.         $SUDO tar -xzf - -C "$OLLAMA_INSTALL_DIR"
  79.     BUNDLE=1
  80.     if [ "$OLLAMA_INSTALL_DIR/bin/ollama" != "$BINDIR/ollama" ] ; then
  81.         status "Making ollama accessible in the PATH in $BINDIR"
  82.         $SUDO ln -sf "$OLLAMA_INSTALL_DIR/ollama" "$BINDIR/ollama"
  83.     fi
  84. else
  85.     status "Downloading Linux ${ARCH} CLI"
  86.     proxychains curl --fail --show-error --location --progress-bar -o "$TEMP_DIR/ollama"\
  87.     "https://ollama.com/download/ollama-linux-${ARCH}${VER_PARAM}"
  88.     $SUDO install -o0 -g0 -m755 $TEMP_DIR/ollama $OLLAMA_INSTALL_DIR/ollama
  89.     BUNDLE=0
  90.     if [ "$OLLAMA_INSTALL_DIR/ollama" != "$BINDIR/ollama" ] ; then
  91.         status "Making ollama accessible in the PATH in $BINDIR"
  92.         $SUDO ln -sf "$OLLAMA_INSTALL_DIR/ollama" "$BINDIR/ollama"
  93.     fi
  94. fi
  96. install_success() {
  97.     status 'The Ollama API is now available at 127.0.0.1:11434.'
  98.     status 'Install complete. Run "ollama" from the command line.'
  99. }
  100. trap install_success EXIT
  102. # Everything from this point onwards is optional.
  104. configure_systemd() {
  105.     if ! id ollama >/dev/null 2>&1; then
  106.         status "Creating ollama user..."
  107.         $SUDO useradd -r -s /bin/false -U -m -d /usr/share/ollama ollama
  108.     fi
  109.     if getent group render >/dev/null 2>&1; then
  110.         status "Adding ollama user to render group..."
  111.         $SUDO usermod -a -G render ollama
  112.     fi
  113.     if getent group video >/dev/null 2>&1; then
  114.         status "Adding ollama user to video group..."
  115.         $SUDO usermod -a -G video ollama
  116.     fi
  118.     status "Adding current user to ollama group..."
  119.     $SUDO usermod -a -G ollama $(whoami)
  121.     status "Creating ollama systemd service..."
  122.     cat <<EOF | $SUDO tee /etc/systemd/system/ollama.service >/dev/null
  123. [Unit]
  124. Description=Ollama Service
  125. After=network-online.target
  127. [Service]
  128. ExecStart=$BINDIR/ollama serve
  129. User=ollama
  130. Group=ollama
  131. Restart=always
  132. RestartSec=3
  133. Environment="PATH=$PATH"
  135. [Install]
  136. WantedBy=default.target
  137. EOF
  138.     SYSTEMCTL_RUNNING="$(systemctl is-system-running || true)"
  139.     case $SYSTEMCTL_RUNNING in
  140.         running|degraded)
  141.             status "Enabling and starting ollama service..."
  142.             $SUDO systemctl daemon-reload
  143.             $SUDO systemctl enable ollama
  145.             start_service() { $SUDO systemctl restart ollama; }
  146.             trap start_service EXIT
  147.             ;;
  148.     esac
  149. }
  151. if available systemctl; then
  152.     configure_systemd
  153. fi
  155. # WSL2 only supports GPUs via nvidia passthrough
  156. # so check for nvidia-smi to determine if GPU is available
  157. if [ "$IS_WSL2" = true ]; then
  158.     if available nvidia-smi && [ -n "$(nvidia-smi | grep -o "CUDA Version: [0-9]*\.[0-9]*")" ]; then
  159.         status "Nvidia GPU detected."
  160.     fi
  161.     install_success
  162.     exit 0
  163. fi
  165. # Install GPU dependencies on Linux
  166. if ! available lspci && ! available lshw; then
  167.     warning "Unable to detect NVIDIA/AMD GPU. Install lspci or lshw to automatically detect and install GPU dependencies."
  168.     exit 0
  169. fi
  171. check_gpu() {
  172.     # Look for devices based on vendor ID for NVIDIA and AMD
  173.     case $1 in
  174.         lspci)
  175.             case $2 in
  176.                 nvidia) available lspci && lspci -d '10de:' | grep -q 'NVIDIA' || return 1 ;;
  177.                 amdgpu) available lspci && lspci -d '1002:' | grep -q 'AMD' || return 1 ;;
  178.             esac ;;
  179.         lshw)
  180.             case $2 in
  181.                 nvidia) available lshw && $SUDO lshw -c display -numeric -disable network | grep -q 'vendor: .* \[10DE\]' || return 1 ;;
  182.                 amdgpu) available lshw && $SUDO lshw -c display -numeric -disable network | grep -q 'vendor: .* \[1002\]' || return 1 ;;
  183.             esac ;;
  184.         nvidia-smi) available nvidia-smi || return 1 ;;
  185.     esac
  186. }
  188. if check_gpu nvidia-smi; then
  189.     status "NVIDIA GPU installed."
  190.     exit 0
  191. fi
  193. if ! check_gpu lspci nvidia && ! check_gpu lshw nvidia && ! check_gpu lspci amdgpu && ! check_gpu lshw amdgpu; then
  194.     install_success
  195.     warning "No NVIDIA/AMD GPU detected. Ollama will run in CPU-only mode."
  196.     exit 0
  197. fi
  199. if check_gpu lspci amdgpu || check_gpu lshw amdgpu; then
  200.     if [ $BUNDLE -ne 0 ]; then
  201.         status "Downloading Linux ROCm ${ARCH} bundle"
  202.         proxychains curl --fail --show-error --location --progress-bar \
  203.             "https://ollama.com/download/ollama-linux-${ARCH}-rocm.tgz${VER_PARAM}" | \
  204.             $SUDO tar -xzf - -C "$OLLAMA_INSTALL_DIR"
  206.         install_success
  207.         status "AMD GPU ready."
  208.         exit 0
  209.     fi
  210.     # Look for pre-existing ROCm v6 before downloading the dependencies
  211.     for search in "${HIP_PATH:-''}" "${ROCM_PATH:-''}" "/opt/rocm" "/usr/lib64"; do
  212.         if [ -n "${search}" ] && [ -e "${search}/libhipblas.so.2" -o -e "${search}/lib/libhipblas.so.2" ]; then
  213.             status "Compatible AMD GPU ROCm library detected at ${search}"
  214.             install_success
  215.             exit 0
  216.         fi
  217.     done
  219.     status "Downloading AMD GPU dependencies..."
  220.     $SUDO rm -rf /usr/share/ollama/lib
  221.     $SUDO chmod o+x /usr/share/ollama
  222.     $SUDO install -o ollama -g ollama -m 755 -d /usr/share/ollama/lib/rocm
  223.     proxychains curl --fail --show-error --location --progress-bar "https://ollama.com/download/ollama-linux-amd64-rocm.tgz${VER_PARAM}" \
  224.         | $SUDO tar zx --owner ollama --group ollama -C /usr/share/ollama/lib/rocm .
  225.     install_success
  226.     status "AMD GPU ready."
  227.     exit 0
  228. fi
  230. CUDA_REPO_ERR_MSG="NVIDIA GPU detected, but your OS and Architecture are not supported by NVIDIA.  Please install the CUDA driver manually https://docs.nvidia.com/cuda/cuda-installation-guide-linux/"
  231. # ref: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#rhel-7-centos-7
  232. # ref: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#rhel-8-rocky-8
  233. # ref: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#rhel-9-rocky-9
  234. # ref: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#fedora
  235. install_cuda_driver_yum() {
  236.     status 'Installing NVIDIA repository...'
  237.     
  238.     case $PACKAGE_MANAGER in
  239.         yum)
  240.             $SUDO $PACKAGE_MANAGER -y install yum-utils
  241.             if proxychains curl -I --silent --fail --location "https://developer.download.nvidia.com/compute/cuda/repos/$1$2/$(uname -m | sed -e 's/aarch64/sbsa/')/cuda-$1$2.repo" >/dev/null ; then
  242.                 $SUDO $PACKAGE_MANAGER-config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/$1$2/$(uname -m | sed -e 's/aarch64/sbsa/')/cuda-$1$2.repo
  243.             else
  244.                 error $CUDA_REPO_ERR_MSG
  245.             fi
  246.             ;;
  247.         dnf)
  248.             if proxychains curl -I --silent --fail --location "https://developer.download.nvidia.com/compute/cuda/repos/$1$2/$(uname -m | sed -e 's/aarch64/sbsa/')/cuda-$1$2.repo" >/dev/null ; then
  249.                 $SUDO $PACKAGE_MANAGER config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/$1$2/$(uname -m | sed -e 's/aarch64/sbsa/')/cuda-$1$2.repo
  250.             else
  251.                 error $CUDA_REPO_ERR_MSG
  252.             fi
  253.             ;;
  254.     esac
  256.     case $1 in
  257.         rhel)
  258.             status 'Installing EPEL repository...'
  259.             # EPEL is required for third-party dependencies such as dkms and libvdpau
  260.             $SUDO $PACKAGE_MANAGER -y install https://dl.fedoraproject.org/pub/epel/epel-release-latest-$2.noarch.rpm || true
  261.             ;;
  262.     esac
  264.     status 'Installing CUDA driver...'
  266.     if [ "$1" = 'centos' ] || [ "$1$2" = 'rhel7' ]; then
  267.         $SUDO $PACKAGE_MANAGER -y install nvidia-driver-latest-dkms
  268.     fi
  270.     $SUDO $PACKAGE_MANAGER -y install cuda-drivers
  271. }
  273. # ref: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#ubuntu
  274. # ref: https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html#debian
  275. install_cuda_driver_apt() {
  276.     status 'Installing NVIDIA repository...'
  277.     if proxychains curl -I --silent --fail --location "https://developer.download.nvidia.com/compute/cuda/repos/$1$2/$(uname -m | sed -e 's/aarch64/sbsa/')/cuda-keyring_1.1-1_all.deb" >/dev/null ; then
  278.         proxychains curl -fsSL -o $TEMP_DIR/cuda-keyring.deb https://developer.download.nvidia.com/compute/cuda/repos/$1$2/$(uname -m | sed -e 's/aarch64/sbsa/')/cuda-keyring_1.1-1_all.deb
  279.     else
  280.         error $CUDA_REPO_ERR_MSG
  281.     fi
  283.     case $1 in
  284.         debian)
  285.             status 'Enabling contrib sources...'
  286.             $SUDO sed 's/main/contrib/' < /etc/apt/sources.list | $SUDO tee /etc/apt/sources.list.d/contrib.list > /dev/null
  287.             if [ -f "/etc/apt/sources.list.d/debian.sources" ]; then
  288.                 $SUDO sed 's/main/contrib/' < /etc/apt/sources.list.d/debian.sources | $SUDO tee /etc/apt/sources.list.d/contrib.sources > /dev/null
  289.             fi
  290.             ;;
  291.     esac
  293.     status 'Installing CUDA driver...'
  294.     $SUDO dpkg -i $TEMP_DIR/cuda-keyring.deb
  295.     $SUDO apt-get update
  297.     [ -n "$SUDO" ] && SUDO_E="$SUDO -E" || SUDO_E=
  298.     DEBIAN_FRONTEND=noninteractive $SUDO_E apt-get -y install cuda-drivers -q
  299. }
  301. if [ ! -f "/etc/os-release" ]; then
  302.     error "Unknown distribution. Skipping CUDA installation."
  303. fi
  305. . /etc/os-release
  307. OS_NAME=$ID
  308. OS_VERSION=$VERSION_ID
  310. PACKAGE_MANAGER=
  311. for PACKAGE_MANAGER in dnf yum apt-get; do
  312.     if available $PACKAGE_MANAGER; then
  313.         break
  314.     fi
  315. done
  317. if [ -z "$PACKAGE_MANAGER" ]; then
  318.     error "Unknown package manager. Skipping CUDA installation."
  319. fi
  321. if ! check_gpu nvidia-smi || [ -z "$(nvidia-smi | grep -o "CUDA Version: [0-9]*\.[0-9]*")" ]; then
  322.     case $OS_NAME in
  323.         centos|rhel) install_cuda_driver_yum 'rhel' $(echo $OS_VERSION | cut -d '.' -f 1) ;;
  324.         rocky) install_cuda_driver_yum 'rhel' $(echo $OS_VERSION | cut -c1) ;;
  325.         fedora) [ $OS_VERSION -lt '39' ] && install_cuda_driver_yum $OS_NAME $OS_VERSION || install_cuda_driver_yum $OS_NAME '39';;
  326.         amzn) install_cuda_driver_yum 'fedora' '37' ;;
  327.         debian) install_cuda_driver_apt $OS_NAME $OS_VERSION ;;
  328.         ubuntu) install_cuda_driver_apt $OS_NAME $(echo $OS_VERSION | sed 's/\.//') ;;
  329.         *) exit ;;
  330.     esac
  331. fi
  333. if ! lsmod | grep -q nvidia || ! lsmod | grep -q nvidia_uvm; then
  334.     KERNEL_RELEASE="$(uname -r)"
  335.     case $OS_NAME in
  336.         rocky) $SUDO $PACKAGE_MANAGER -y install kernel-devel kernel-headers ;;
  337.         centos|rhel|amzn) $SUDO $PACKAGE_MANAGER -y install kernel-devel-$KERNEL_RELEASE kernel-headers-$KERNEL_RELEASE ;;
  338.         fedora) $SUDO $PACKAGE_MANAGER -y install kernel-devel-$KERNEL_RELEASE ;;
  339.         debian|ubuntu) $SUDO apt-get -y install linux-headers-$KERNEL_RELEASE ;;
  340.         *) exit ;;
  341.     esac
  343.     NVIDIA_CUDA_VERSION=$($SUDO dkms status | awk -F: '/added/ { print $1 }')
  344.     if [ -n "$NVIDIA_CUDA_VERSION" ]; then
  345.         $SUDO dkms install $NVIDIA_CUDA_VERSION
  346.     fi
  348.     if lsmod | grep -q nouveau; then
  349.         status 'Reboot to complete NVIDIA CUDA driver install.'
  350.         exit 0
  351.     fi
  353.     $SUDO modprobe nvidia
  354.     $SUDO modprobe nvidia_uvm
  355. fi
  357. # make sure the NVIDIA modules are loaded on boot with nvidia-persistenced
  358. if available nvidia-persistenced; then
  359.     $SUDO touch /etc/modules-load.d/nvidia.conf
  360.     MODULES="nvidia nvidia-uvm"
  361.     for MODULE in $MODULES; do
  362.         if ! grep -qxF "$MODULE" /etc/modules-load.d/nvidia.conf; then
  363.             echo "$MODULE" | $SUDO tee -a /etc/modules-load.d/nvidia.conf > /dev/null
  364.         fi
  365.     done
  366. fi
  368. status "NVIDIA GPU ready."
  369. install_success

image

  • 手动安装参考：

配置环境变量

登录后复制

  1. vim /home/viadmin/.bashrc
  2. export OLLAMA_HOST=http://10.10.16.60:11434
  3. systemctl set-environment OLLAMA_HOST=http://10.10.16.60:11434
  4. source .bashrc
  5. systemctl restart ollama
  6. systemctl status ollama

image

image

Ollama相关参数

登录后复制

  1. viadmin@ollama-pro:~$ ollama --help
  2. Large language model runner
  4. Usage:
  5.   ollama [flags]
  6.   ollama [command]
  8. Available Commands:
  9.   serve       Start ollama
  10.   create      Create a model from a Modelfile
  11.   show        Show information for a model
  12.   run         Run a model
  13.   stop        Stop a running model
  14.   pull        Pull a model from a registry
  15.   push        Push a model to a registry
  16.   list        List models
  17.   ps          List running models
  18.   cp          Copy a model
  19.   rm          Remove a model
  20.   help        Help about any command
  22. Flags:
  23.   -h, --help      help for ollama
  24.   -v, --version   Show version information
  26. Use "ollama [command] --help" for more information about a command.

安装模型报错问题

image

登录后复制

  1. Error: could not connect to ollama app, is it running?

  • 经过查询，发现是需要先启动ollama app，启动方式是：sudo ollama serve

image

  • 上述启动是一个交互式的形式，可以使用screen命令，进入此空间后再执行。
  • 上述执行完成之后就可以下载模型了。
  • https://github.com/ollama/ollama

image

image

最终效果

image

使用docker安装Open-Webui

官方文档

登录后复制

  1. 官方：
  2. sudo docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
  4. 自定义安装将3000端口改为80端口
  5. docker run -d -p 80:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

image

  • 经过测试上述操作，虽然open-webui成功启动，但是无法识别本地安装的ollama模型，所以需要使用以下方式启动

登录后复制

  1. sudo docker run -d --network=host -e OLLAMA_BASE_URL=http://127.0.0.1:11434 -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main

效果展示

image

Docker搭建Dify

  1. 克隆 Dify 源代码至本地环境。git clone https://github.com/langgenius/dify.git
  2. 进入 Dify 源代码的 Docker 目录cd dify/docker
  3. 复制环境配置文件cp .env.example .env
  4. 启动 Docker 容器

  • 根据你系统上的 Docker Compose 版本，选择合适的命令来启动容器。你可以通过 $ docker compose version 命令检查版本，详细说明请参考 Docker 官方文档：

  • 如果版本是 Docker Compose V2，使用以下命令：docker compose up -d

  • 如果版本是 Docker Compose V1，使用以下命令：docker-compose up -d

  • 最后检查是否所有容器都正常运行：docker compose ps

  • 参考：

  • https://docs.dify.ai/zh-hans/getting-started/install-self-hosted/docker-compose

  • 上述完成之后，正常访问部署到服务器的IP地址，然后会让你初始化设置账户和密码，然后登录进去下面是登录进去的效果。

image

在Dify中加载Ollama

  • 这里刚开始踩了坑，就是默认情况下新版本的Ollama安装成功之后就是以Systemd服务的形式启动，而这个刚开始我没仔细看，导致自己也手动使用ollama serve启动了服务，这样导致互相混淆了，最终导致的结果就是我无论怎么配置加载，在Dify中都无法查看到Ollama所加载的模型。
  • 解决办法就是我关停了自己使用ollama serve启动的服务，然后使用Systemd形式进行管理配置服务，并且配置侦听所有IP地址。
  • 由于上述的操作，导致对应下载的模型路径不对，所以需要更改模型的路径。

需要添加到Systemd的配置参数

  • sudo vim /etc/systemd/system/ollama.service

登录后复制

  1. Environment="OLLAMA_HOST=0.0.0.0"
  2. Environment="OLLAMA_MODELS=/home/viadmin/.ollama/models"

  • 最终的配置结果

登录后复制

  1. viadmin@ollama-pro:/etc/systemd/system$ cat ollama.service 
  2. [Unit]
  3. Description=Ollama Service
  4. After=network-online.target
  6. [Service]
  7. Environment="OLLAMA_HOST=0.0.0.0"
  8. ExecStart=/usr/local/bin/ollama serve
  9. User=ollama
  10. Group=ollama
  11. Restart=always
  12. RestartSec=3
  13. Environment="PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/snap/bin"
  14. Environment="OLLAMA_MODELS=/home/viadmin/.ollama/models"
  16. [Install]
  17. WantedBy=default.target

  • 目录权限记得给够，不行就给777

  • 重启服务

登录后复制

  1. sudo systemctl daemon-reload
  2. sudo systemctl restart ollama.service
  3. sudo systemctl status ollama.service

image

image

迷茫的人生，需要不断努力，才能看清远方模糊的志向！

标签： ubuntu linux 运维
本文转载自: https://blog.csdn.net/qq_35485206/article/details/143459935
版权归原作者 心上之秋 所有， 如有侵权，请联系我们删除。
登录后发布评论

“国内网络在Ubuntu 22.04中在线安装Ollama并配置Open-Webui&Dify”的评论:

还没有评论
关于作者
...
overfit同步小助手
文章同步
相关阅读

网络安全法-网络运行安全

使用selenium/drissionpage时如何阻止chrome自动跳转http到https

docker desktop 里部署的Open WebUI 管理员密码忘记了的处理方法

在ubuntu20.04中搭建onsite比赛运行环境

利用开源的低代码表单设计器FcDesigner高效管理和渲染复杂表单结构

Kafka学习笔记

【前端】浏览器输入url到页面呈现发生了什么？

文章导航
没有找到文章内容分段。。。