在UBUNTU内使用DOCKER配置PYTHON项目环境

一、创建requirements.txt

torch>=1.4.0
torchvision>=0.5.0
dominate>=2.4.0
visdom>=0.1.8.8
wandb

二、创建Dockerfile

FROM nvidia/cuda:12.1.1-cudnn8-devel-ubuntu22.04
MAINTAINER Ma Yue

ENV LANG=C.UTF-8 LC_ALL=C.UTF-8

RUN sed -i 's#http://archive.ubuntu.com/#http://mirrors.aliyun.com/#' /etc/apt/sources.list \
    && sed -i 's#http://security.ubuntu.com/#http://mirrors.aliyun.com/#' /etc/apt/sources.list


#RUN apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
RUN apt-get update \
    && apt-get install -y software-properties-common curl libgl1-mesa-glx libglib2.0-0 libsm6 libxrender1 libxext-dev scrot vim\
    && add-apt-repository ppa:deadsnakes/ppa \
    && apt-get remove -y software-properties-common \
    && apt autoremove -y \
    && apt-get update \
    && apt-get install -y python3.8 python3.8-dev python3.8-distutils\
    && curl -o /tmp/get-pip.py "https://bootstrap.pypa.io/get-pip.py" \
    && python3.8 /tmp/get-pip.py \
    && apt-get remove -y curl \
    && apt autoremove -y \
    && rm -rf /var/lib/apt/lists/*
COPY requirements.txt /requirements.txt
RUN ln -sfn /usr/bin/python3.8 /usr/bin/python3 && ln -sfn /usr/bin/python3 /usr/bin/python && ln -sfn /usr/bin/pip3 /usr/bin/pip
RUN pip install --no-cache-dir -r /requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple

FROM nvidia/cuda:12.1.1-cudnn8-devel-ubuntu22.04
MAINTAINER Ma Yue

ENV LANG=C.UTF-8 LC_ALL=C.UTF-8

#定义时区参数
ENV TZ=Asia/Shanghai
#设置时区
RUN ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo '$TZ' > /etc/timezone


RUN sed -i 's#http://archive.ubuntu.com/#http://mirrors.aliyun.com/#' /etc/apt/sources.list \
    && sed -i 's#http://security.ubuntu.com/#http://mirrors.aliyun.com/#' /etc/apt/sources.list


#RUN apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
RUN apt-get update \
    && apt-get install -y software-properties-common curl libgl1-mesa-glx libglib2.0-0 libsm6 libxrender1 libxext-dev scrot vim\
    && add-apt-repository ppa:deadsnakes/ppa \
    && apt-get remove -y software-properties-common \
    && apt autoremove -y \
    && apt-get update \
    && apt-get install -y python3.8 python3.8-dev python3.8-distutils\
    && curl -o /tmp/get-pip.py "https://bootstrap.pypa.io/get-pip.py" \
    && python3.8 /tmp/get-pip.py \
    && apt-get remove -y curl \
    && apt autoremove -y \
    && rm -rf /var/lib/apt/lists/*
COPY requirements.txt /requirements.txt
RUN ln -sfn /usr/bin/python3.8 /usr/bin/python3 && ln -sfn /usr/bin/python3 /usr/bin/python && ln -sfn /usr/bin/pip3 /usr/bin/pip
RUN pip install --no-cache-dir -r /requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple

三、配置DOCKER

先进入root模式：sudo su

输入密码：[sudo] studentfive 的密码：

进入项目目录

root@admin:

cd /date1/sun/docker

得到：Dockerfile  requirements.txt

root@admin:/date1/sun/docker# docker build -t cyclegan:1.0 .

docker build -t cyclegan:1.0 .

docker build --network host -t cyclegan:1.0 .

build失败了，查看一下映射是否成功： docker images

没有项目，继续build：docker build -t cyclegan:1.0 .

build成功了

进入docker镜像：docker images  

映射项目文件夹：

docker run --name cyclegan -v /date1/sun/cyclegan/pytorch-CycleGAN-and-pix2pix-master:/date1/sun/cyclegan/pytorch-CycleGAN-and-pix2pix-master -it -d --gpus all -e PYTHONIOENCODING=UTF-8 cyclegan:1.0 /bin/bash

得到：24607a3542cae01f14817931916127acc348995618591190e37d975cdcb9c757

查看docker: docker ps -a

另外，如果需要使用visdom，则需在创建docker时映射端口：

docker run --name cyclegan -v /date1/sun/cyclegan/pytorch-CycleGAN-and-pix2pix-master:/date1/sun/cyclegan/pytorch-CycleGAN-and-pix2pix-master -p 8888:8097 -it -d -e PYTHONIOENCODING=UTF-8 --gpus all cyclegan:1.0 /bin/bash

其中，8097是docker中visdom的端口，8888是自己创建的在ubuntu中的端口

四、在docker中项目开发

进docker：

docker exec -it cyclegan /bin/bash

进项目：

cd /date1/sun/cyclegan/pytorch-CycleGAN-and-pix2pix-master/

然后进行项目开发，例如运行：

python train.py --dataroot ./datasets/outline2micro --name outline2micro_cyclegan --model cycle_gan

测试：

python test.py --dataroot ./datasets/outline2micro --name outline2micro_cyclegan --model cycle_gan

结束之后退出docker：exit

五、停止、删除所有的docker容器和镜像
列出所有的容器 ID

docker ps -a

停止所有的容器

docker stop $(docker ps -aq)

删除所有的容器

docker rm $(docker ps -aq)

删除所有的镜像

docker rmi $(docker images -q)

六、

print(torch.__version__)

torch.cuda.is_available()

七、启动visdom

python -m visdom.server

1.如果报错，

Checking for scripts.
Downloading scripts, this may take a little while
ERROR:root:Error [Errno -3] Temporary failure in name resolution while downloading https://unpkg.com/jquery@3.1.1/dist/jquery.min.js
ERROR:root:Error [Errno -3] Temporary failure in name resolution while downloading https://unpkg.com/bootstrap@3.3.7/dist/js/bootstrap.min.js
ERROR:root:Error [Errno -3] Temporary failure in name resolution while downloading https://unpkg.com/react@16.2.0/umd/react.production.min.js

解决方法：

进入docker后，输入

cd /usr/local/lib/python3.8/dist-packages/visdom/server

vim run_server.py

按i，进入Insert模式，下到最下面将这一行 download_scripts()注释掉，

def download_scripts_and_run():
   # download_scripts()
    main()

按ESC键，输入:wq!退出

2.如果visdom网页蓝屏，且终端报错

ERROR:tornado.general:Could not open static file '/usr/local/lib/python3.8/dist-packages/visdom/static/js/plotly-plotly.min.js'
ERROR:tornado.general:Could not open static file '/usr/local/lib/python3.8/dist-packages/visdom/static/js/d3.v3.min.js'
ERROR:tornado.general:Could not open static file '/usr/local/lib/python3.8/dist-packages/visdom/static/js/d3-selection-multi.v1.js'
ERROR:tornado.general:Could not open static file '/usr/local/lib/python3.8/dist-packages/visdom/static/js/saveSvgAsPng.js'
ERROR:tornado.general:Could not open static file '/usr/local/lib/python3.8/dist-packages/visdom/user/style.css'
INFO:tornado.access:200 GET / (172.17.0.1) 20.92ms
WARNING:tornado.access:404 GET /static/css/bootstrap.min.css (172.17.0.1) 2.68ms
WARNING:tornado.access:404 GET /static/js/jquery.min.js (172.17.0.1) 1.36ms
WARNING:tornado.access:404 GET /static/js/bootstrap.min.js (172.17.0.1) 1.19ms
WARNING:tornado.access:404 GET /static/css/react-resizable-styles.css (172.17.0.1) 1.31ms
WARNING:tornado.access:404 GET /static/css/react-grid-layout-styles.css (172.17.0.1) 1.12ms
WARNING:tornado.access:404 GET /static/js/react-react.min.js (172.17.0.1) 1.89ms
WARNING:tornado.access:404 GET /static/js/react-dom.min.js (172.17.0.1) 2.23ms
WARNING:tornado.access:404 GET /static/js/layout_bin_packer.js (172.17.0.1) 2.55ms
 

解决方法：

 先删除容器里的static文件夹，

cd /usr/local/lib/python3.8/dist-packages/visdom

ls看一下，应该存在static文件夹

输入：

rm -r static

ls看一下，应该已经删除了static文件夹

输入exit退出docker模式，自动进入root模式

然后在root@admin下执行命令：

docker cp /date1/sun/docker/static cyclegan:/usr/local/lib/python3.8/dist-packages/visdom

watch -n 0.1 nvidia-smi

https://www.cnblogs.com/2205254761qq/p/11863928.html

bashdistribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update
sudo apt-get install -y nvidia-docker2

bash sudo systemctl restart docker

服务器重启之后，找不懂到mage, docker，

systemctl status docker

1.当docker状态是running

先重新拉个镜像（不确定这步可不可以省略），之后跟着链接里的走。

服务器重启之后，docker ps 看不到任何容器_重启容器服务 linux-CSDN博客

内容为：

重启服务器之后，docker服务没有做自启动·

docker ps

报错：Cannot connect to the Docker daemon at unix:///var/run/docker.sock. Is the docker daemon running?

因为没启动，需要手动重启

systemctl start docker

然后docker ps  是没有任何容器启动的

所以要查询所有容器的id

docker container ls -a

这时候你能看到所有你之前被关掉的容器

start一下，就可以了

docker container start 01738895328c（container Id）

stop同上了。希望对你有帮助。

自动重启：

docker update --restart=always NAME（容器的名称）

————————————————

                            版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
                        
原文链接：https://blog.csdn.net/FBX_fbx_FBX/article/details/109570694

2. 当docker状态是failed

failed to start daemon: Error initializing network controller: Error creating default “bridge“ ***-CSDN博客

输入，

sudo dockerd --debug

得到报错，

failed to start daemon: Error initializing network controller: error creating default "bridge" network: cannot create network ee4e7f4c5317b3dec2b6fe5db0044ba535c038e4282709f5fd4bce3f3fe4638f (docker0): conflicts with network c9e21497a9d4268189fe6a0b16279efb81d45bce8564104aa983ab1570f03659 (docker0): networks have same bridge name

解决方法，

sudo rm -rf /var/lib/docker/network
sudo systemctl start docker

1. 如果还是报错 ；

sudo systemctl stop firewalld
sudo systemctl start docker
sudo systemctl status docker

之后，同上

systemctl start docker
然后docker ps  是没有任何容器启动的

所以要查询所有容器的id

docker container ls -a
这时候你能看到所有你之前被关掉的容器

start一下，就可以了

docker container start 01738895328c（container Id）

3. 连接服务器蓝屏/黑屏

方法1：重启xrdp服务 （个人推荐）

sudo  /etc/init.d/xrdp  restart

之后使用账号密码成功登陆：

xrdp远程桌面蓝/黑屏问题解决方法_xrdp远程桌面进不去界面-CSDN博客

4.

启用 Docker 服务开机自启动

为了确保 Docker 服务在系统重启后自动启动，可以启用它：

sudo systemctl enable docker