欧拉服务器操作系统部署deekseep（Ollama+DeekSeep+open WebUI）

​​一、解压并安装 Ollama​​

1. 解压文件

# 1. 解压文件（默认会得到一个二进制文件）
[root@localhost ~]# tar -xzvf ollama-linux-amd64.tgz 
bin/ollama
lib/ollama/cuda_v11/
lib/ollama/cuda_v11/libggml-cuda.so
lib/ollama/cuda_v11/libcublas.so.11
lib/ollama/cuda_v11/libcublas.so.11.5.1.109
lib/ollama/cuda_v11/libcublasLt.so.11.5.1.109
lib/ollama/cuda_v11/libcudart.so.11.3.109
lib/ollama/cuda_v11/libcublasLt.so.11
lib/ollama/cuda_v11/libcudart.so.11.0
lib/ollama/cuda_v12/
lib/ollama/cuda_v12/libggml-cuda.so
lib/ollama/cuda_v12/libcudart.so.12
lib/ollama/cuda_v12/libcudart.so.12.8.90
lib/ollama/cuda_v12/libcublasLt.so.12
lib/ollama/cuda_v12/libcublas.so.12.8.4.1
lib/ollama/cuda_v12/libcublas.so.12
lib/ollama/cuda_v12/libcublasLt.so.12.8.4.1
lib/ollama/libggml-base.so
lib/ollama/libggml-cpu-alderlake.so
lib/ollama/libggml-cpu-haswell.so
lib/ollama/libggml-cpu-icelake.so
lib/ollama/libggml-cpu-sandybridge.so
lib/ollama/libggml-cpu-skylakex.so

2. 创建安装目录​

[root@localhost ~]# sudo mkdir -p /opt/ollama
[root@localhost ~]# sudo cp -r bin/ /opt/ollama/
[root@localhost ~]# sudo cp -r lib/ /opt/ollama/

3.创建符号链接到系统路径

sudo ln -s /opt/ollama/bin/ollama /usr/local/bin/ollama

4. 配置动态库路径​​

确保 Ollama 能找到 CUDA 和 CPU 架构的动态库：

[root@localhost ~]# echo "export LD_LIBRARY_PATH=/opt/ollama/lib/ollama:\$LD_LIBRARY_PATH" | sudo tee -a /etc/profile.d/ollama.sh
export LD_LIBRARY_PATH=/opt/ollama/lib/ollama:$LD_LIBRARY_PATH
[root@localhost ~]# source /etc/profile.d/ollama.sh

5. 验证安装

提示无法连接到正在运行的Ollama实例,是因为没有配置service服务

[root@localhost ~]# ollama --version
Warning: could not connect to a running Ollama instance
Warning: client version is 0.6.5

链接：ollama-linux-amd64.tgz下载地址

二、配置 Ollama 服务​​

1. ​​创建专用用户和目录​

# 创建 ollama 用户的主目录（如果不存在）
sudo mkdir -p /home/ollama
# 设置目录所有权为 ollama 用户
sudo chown ollama:ollama /home/ollama
# 确保目录权限正确（用户可读写）
sudo chmod 700 /home/ollama
sudo useradd -r -s /bin/false ollama
#创建模型存储目录并授权​
sudo mkdir -p /etc/ollama /var/lib/ollama
sudo chown -R ollama:ollama /etc/ollama /var/lib/ollama

2. ​​配置 Systemd 服务​​

sudo tee /etc/systemd/system/ollama.service <<'EOF'
[Unit]
Description=Ollama Service
After=network-online.target
Requires=network-online.target[Service]
User=ollama
Group=ollama
ExecStart=/usr/local/bin/ollama serve
Restart=on-failure# 环境变量
Environment="PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin"
Environment="OLLAMA_MODELS=/var/lib/ollama"
Environment="OLLAMA_HOST=0.0.0.0:11434"  # 绑定所有接口，端口 11434# 资源限制
MemoryLimit=4G
CPUQuota=200%# 安全配置
NoNewPrivileges=true
PrivateTmp=true
ProtectSystem=full[Install]
WantedBy=multi-user.target
EOF

3. ​​启动服务

sudo systemctl daemon-reload
sudo systemctl start ollama
sudo systemctl enable ollama

4. 放行防火墙端口

sudo firewall-cmd --permanent --add-port=11434/tcp
sudo firewall-cmd --reload
[root@koji-builder ~]# sudo firewall-cmd --list-ports | grep 11434
11434/tcp

三、验证安装​​

1. ​​检查服务状态​​

[root@koji-builder ~]# sudo systemctl status ollama  
● ollama.service - Ollama ServiceLoaded: loaded (/etc/systemd/system/ollama.service; enabled; vendor preset: disabled)Active: active (running) since Mon 2025-04-14 11:41:24 CST; 7min agoMain PID: 357650 (ollama)Tasks: 16 (limit: 100784)Memory: 17.8MCGroup: /system.slice/ollama.service└─357650 /usr/local/bin/ollama serve4月 14 11:41:24 koji-builder systemd[1]: Started Ollama Service.
4月 14 11:41:24 koji-builder ollama[357650]: 2025/04/14 11:41:24 routes.go:1231: INFO server config env="map[CUDA_VISIBLE_DEVICES: GPU_DEVICE_ORDINAL:>
4月 14 11:41:24 koji-builder ollama[357650]: time=2025-04-14T11:41:24.960+08:00 level=INFO source=images.go:458 msg="total blobs: 0"
4月 14 11:41:24 koji-builder ollama[357650]: time=2025-04-14T11:41:24.960+08:00 level=INFO source=images.go:465 msg="total unused blobs removed: 0"
4月 14 11:41:24 koji-builder ollama[357650]: time=2025-04-14T11:41:24.960+08:00 level=INFO source=routes.go:1298 msg="Listening on 127.0.0.1:11434 (ve>
4月 14 11:41:24 koji-builder ollama[357650]: time=2025-04-14T11:41:24.961+08:00 level=INFO source=gpu.go:217 msg="looking for compatible GPUs"
4月 14 11:41:24 koji-builder ollama[357650]: time=2025-04-14T11:41:24.974+08:00 level=INFO source=gpu.go:377 msg="no compatible GPUs were discovered"
4月 14 11:41:24 koji-builder ollama[357650]: time=2025-04-14T11:41:24.974+08:00 level=INFO source=types.go:130 msg="inference compute" id=0 library=cp>
4月 14 11:47:58 koji-builder ollama[357650]: [GIN] 2025/04/14 - 11:47:58 | 200 |      99.397µs |       127.0.0.1 | GET      "/"
lines 1-18/18 (END)

​​2. 验证端口监听状态​​

[root@koji-builder ~]# ss -tulnp | grep 11434
tcp   LISTEN 0      128                 *:11434            *:*    users:(("ollama",pid=369774,fd=3))                        
[root@koji-builder ~]# 

3. ​​测试 API 连通性​​

[root@koji-builder ~]# curl http://localhost:11434
Ollama is running

四、安装 DeepSeek 模型​​

1. ​​直接拉取模型（需网络通畅）​

[root@koji-builder ~]# ollama pull deepseek-r1:7b
pulling manifest 
pulling 96c415656d37... 100% ▕███████████████████████████████████████████████████████████████████████████████████████▏ 4.7 GB                         
pulling 369ca498f347... 100% ▕███████████████████████████████████████████████████████████████████████████████████████▏  387 B                         
pulling 6e4c38e1172f... 100% ▕███████████████████████████████████████████████████████████████████████████████████████▏ 1.1 KB                         
pulling f4d24e9138dd... 100% ▕███████████████████████████████████████████████████████████████████████████████████████▏  148 B                         
pulling 40fb844194b2... 100% ▕███████████████████████████████████████████████████████████████████████████████████████▏  487 B                         
verifying sha256 digest 
writing manifest 
success 
[root@koji-builder ~]# 

2. ​​运行模型​

[root@koji-builder ~]# ollama run deepseek-r1:7b
>>> 你好，DeepSeek！
<think></think>你好！我是DeepSeek-R1，一个由深度求索公司开发的智能助手，我会尽我所能为您提供帮助。请问有什么可以为您服务的？

五、安装 Open WebUI

1.添加 Docker 仓库

sudo dnf config-manager --add-repo=https://repo.huaweicloud.com/docker-ce/linux/centos/docker-ce.repo
sudo sed -i 's+download.docker.com+repo.huaweicloud.com/docker-ce+' /etc/yum.repos.d/docker-ce.repo
sudo sed -i 's+\$releasever+8+' /etc/yum.repos.d/docker-ce.repo

2.安装 Docker

使用以下命令安装 Docker：

yum remove runc
sudo yum install docker-ce docker-ce-cli containerd.io -y

3.启动并启用 Docker 服务

启动 Docker 服务并设置为开机自启动：

[root@koji-builder ~]# sudo systemctl start docker
[root@koji-builder ~]# sudo systemctl enable docker --now
[root@koji-builder ~]# systemctl status docker.service 
● docker.service - Docker Application Container EngineLoaded: loaded (/usr/lib/systemd/system/docker.service; enabled; vendor preset: disabled)Active: active (running) since Mon 2025-04-14 14:23:11 CST; 13min agoDocs: https://docs.docker.comMain PID: 366089 (dockerd)Tasks: 52Memory: 4.8GCGroup: /system.slice/docker.service├─366089 /usr/bin/dockerd -H fd:// --containerd=/run/containerd/containerd.sock├─367006 /usr/bin/docker-proxy -proto tcp -host-ip 0.0.0.0 -host-port 3000 -container-ip 172.17.0.2 -container-port 8080└─367013 /usr/bin/docker-proxy -proto tcp -host-ip :: -host-port 3000 -container-ip 172.17.0.2 -container-port 80804月 14 14:23:09 koji-builder dockerd[366089]: time="2025-04-14T14:23:09.876058935+08:00" level=info msg="Loading containers: start."
4月 14 14:23:11 koji-builder dockerd[366089]: time="2025-04-14T14:23:11.681953947+08:00" level=info msg="Firewalld: interface docker0 already part of >
4月 14 14:23:11 koji-builder dockerd[366089]: time="2025-04-14T14:23:11.883178595+08:00" level=info msg="Loading containers: done."
4月 14 14:23:11 koji-builder dockerd[366089]: time="2025-04-14T14:23:11.912847572+08:00" level=info msg="Docker daemon" commit=8e96db1 containerd-snap>
4月 14 14:23:11 koji-builder dockerd[366089]: time="2025-04-14T14:23:11.913161898+08:00" level=info msg="Daemon has completed initialization"
4月 14 14:23:11 koji-builder dockerd[366089]: time="2025-04-14T14:23:11.956811112+08:00" level=info msg="API listen on /run/docker.sock"
4月 14 14:23:11 koji-builder systemd[1]: Started Docker Application Container Engine.
4月 14 14:30:56 koji-builder dockerd[366089]: time="2025-04-14T14:30:56.087596330+08:00" level=error msg="Not continuing with pull after error: contex>
4月 14 14:33:36 koji-builder dockerd[366089]: time="2025-04-14T14:33:36.929237215+08:00" level=info msg="Pull session cancelled"
4月 14 14:33:37 koji-builder dockerd[366089]: time="2025-04-14T14:33:37.845187027+08:00" level=error msg="Not continuing with pull after error: contex>

4.验证 Docker 安装

运行以下命令检查 Docker 版本：

[root@koji-builder ~]# docker --version
Docker version 26.1.3, build b72abbb
[root@koji-builder ~]# 

5.通过 Docker 快速安装​open-webui

[root@koji-builder ~]# docker run -d -p 3000:8080 -e OLLAMA_BASE_URL=http://172.16.104.203:11434 -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
Unable to find image 'ghcr.io/open-webui/open-webui:main' locally
main: Pulling from open-webui/open-webui
8a628cdd7ccc: Pull complete 
2a47a8c4fd5c: Pull complete 
9c538fc35491: Pull complete 
47bbb0afa7fe: Pull complete 
782acb99e453: Pull complete 
4f4fb700ef54: Pull complete 
b7915144f9c7: Pull complete 
633be3c6bab0: Pull complete 
266a80d83771: Pull complete 
7b4a3fa111d1: Pull complete 
693caf783e3a: Pull complete 
dc27e7f6d410: Pull complete 
734b72bc705c: Pull complete 
1f468c116822: Pull complete 
9cbf3744ebb9: Pull complete 
Digest: sha256:ae21d77f8d3c5ab06cdd79b5e1462ef57ca162de04e2ccb8977615e301fcb42a
Status: Downloaded newer image for ghcr.io/open-webui/open-webui:main
e27380dce3851bd9be09ce050bb8b0ffd0054abefe97ab4e3ede61f1989cd100

六、访问open-webui

如果第一次访问ip时没有报500，就可以登录进去关闭openai的API

​​1. 查看 Docker 映射到宿主机webui.db位置

为了解决 OpenWebUI 报错 500

​​1. 查看容器挂载详情​

[root@koji-builder ~]# docker inspect open-webui | grep -A 10 "Mounts""Mounts": [{"Type": "volume","Name": "open-webui","Source": "/var/lib/docker/volumes/open-webui/_data","Destination": "/app/backend/data","Driver": "local","Mode": "z","RW": true,"Propagation": ""}
[root@koji-builder ~]# 

2.定位文件位置

[root@koji-builder ~]# cd /var/lib/docker/volumes/open-webui/_data
[root@koji-builder _data]# ls
cache  uploads  vector_db  webui.db 
[root@koji-builder _data]# 

​​2. 修改修改 config 表中 openai.enable 字段

如果不修改的话，会报500，因为openwebui使用了openai的api，如果不能访问外网的情况下，需要修改数据库表的openai.enable 字段

1.安装sqlite

yum install sqlite

2.备份数据库和停止 Open WebUI 容器​

sudo cp webui.db webui.db.bak
#停止 Open WebUI 容器​
docker stop open-webui

3. 进入 SQLite 交互模式​

[root@koji-builder _data]# sqlite3  webui.db 
SQLite version 3.26.0 2018-12-01 12:34:55
Enter ".help" for usage hints.
sqlite> 

4. 查看所有表，是否有config 表

sqlite> .tables 
alembic_version   config            group             model           
auth              document          knowledge         prompt          
channel           feedback          memory            tag             
channel_member    file              message           tool            
chat              folder            message_reaction  user            
chatidtag         function          migratehistory  
sqlite> 

5. 修改config 表"openai":{"enable"是的修改true

#修改
sqlite> UPDATE config ...> SET data = json_set(...>   data, ...>   '$.openai.enable', ...>   json('false')  ...> ) ...> WHERE id = 1;
#-- 验证结果
sqlite> SELECT json_extract(data, '$.openai.enable') FROM config WHERE id = 1;
0

6. 查看config 表中的"openai":{"enable"是的修改false

sqlite> select * from config;
1|{"version":0,"ui":{"default_locale":"","prompt_suggestions":[{"title":["Help me study","vocabulary for a college entrance exam"],"content":"Help me study vocabulary: write a sentence for me to fill in the blank, and I'll try to pick the correct option."},{"title":["Give me ideas","for what to do with my kids' art"],"content":"What are 5 creative things I could do with my kids' art? I don't want to throw them away, but it's also so much clutter."},{"title":["Tell me a fun fact","about the Roman Empire"],"content":"Tell me a random fun fact about the Roman Empire"},{"title":["Show me a code snippet","of a website's sticky header"],"content":"Show me a code snippet of a website's sticky header in CSS and JavaScript."},{"title":["Explain options trading","if I'm familiar with buying and selling stocks"],"content":"Explain options trading in simple terms if I'm familiar with buying and selling stocks."},{"title":["Overcome procrastination","give me tips"],"content":"Could you start by asking me about instances when I procrastinate the most and then give me some suggestions to overcome it?"},{"title":["Grammar check","rewrite it for better readability "],"content":"Check the following sentence for grammar and clarity: \"[sentence]\". Rewrite it for better readability while maintaining its original meaning."}],"enable_signup":false},"ollama":{"enable":true,"base_urls":["http://172.16.104.203:11434"],"api_configs":{"0":{}}},"openai":{"enable":false,"api_base_urls":["https://api.openai.com/v1"],"api_keys":[""],"api_configs":{"0":{}}}}|0|2025-04-14 08:07:08|2025-04-14 08:50:11.453672
sqlite> 
#-- 退出
sqlite> .quit

7. 重启open-webui容器

[root@koji-builder _data]# docker start  open-webui 
open-webui

如果启动后没有反应需要进容器里启动
[root@koji-builder _data]# docker  exec -it open-webui /bin/bash
root@69d9b2ad6b44:/app/backend# ls
data  dev.sh  open_webui  requirements.txt  start.sh  start_windows.bat
root@69d9b2ad6b44:/app/backend# ./start.sh 

​​3. 访问open-webui

http://ip:3000

七、删除 Docker 上所有与 Open WebUI 相关资源的完整步骤：

​​1. 停止并删除容器​​

# 查找容器名称或 ID
docker ps -a | grep open-webui
# 停止并删除容器（替换为实际容器 ID 或名称）
docker stop open-webui && docker rm open-webui

​​2. 删除数据卷​​

# 查看数据卷列表
docker volume ls | grep open-webui
# 删除数据卷（替换为实际卷名）
docker volume rm open-webui

​​3. 清理镜像​​

# 查看镜像 ID
docker images | grep open-webui
# 删除镜像（替换为实际镜像 ID）
docker rmi ghcr.io/open-webui/open-webui:main

​​4. 强制清理所有残留资源​​

# 删除所有未使用的容器、镜像、卷和网络
docker system prune -af

​​5. 验证清理结果​​

# 检查容器、卷、镜像是否已删除
docker ps -a | grep open-webui      # 应无输出
docker volume ls | grep open-webui  # 应无输出
docker images | grep open-webui     # 应无输出

​​6. 手动清理宿主机的残留文件（可选）​​

# 删除可能残留的本地文件（路径根据实际情况调整）
sudo rm -rf /var/lib/docker/volumes/open-webui