Docker: 小白之路九(从0搭建自己的Docker环境centos7)

Docker环境配置centos7版本

一. 配置对应的docker环境和nvidia-docker(注意加速镜像设置)

二. 环境配置

1. 拉取对应的镜像

docker pull nvidia/cuda:10.2-cudnn7-devel-centos7
1

2. 创建新版本的容器并搭建对应的环境

docker run -it --name felaim_sever_centos7 nvidia/cuda:10.2-cudnn7-devel-centos7 /bin/bash
1

1. 更新对应的源和软件

yum  update
yum upgrade
1
2

2. 安装wget

yum -y install wget
1

3. 确定g++和gcc版本

最好是用5.4.0版本的,4.8.5对有些软件安装会有问题例如OpenCV4.2的GPU版本

gcc --version
1

a. 安装解压缩包bzip2

yum install bzip2
1

b. 下载对应源码并解压(网容易不好,要多试几次才行)

cd /usr/local/src
wget https://ftp.gnu.org/gnu/gcc/gcc-5.4.0/gcc-5.4.0.tar.bz2
tar -jxvf gcc-5.4.0.tar.bz2
1
2
3

c.下载相关依赖组件(网不好,用热点会好一些)

cd gcc-5.4.0
./contrib/download_prerequisites
1
2

d. 建立一个文件夹存放编译的文件

mkdir gcc-build-5.4.0
cd gcc-build-5.4.0
/usr/local/src/gcc-5.4.0/configure --enable-checking=release --enable-languages=c,c++ --disable-multilib  
1
2
3

e. 编译并安装

出现问题,没有安装make

make -j4 //开四个线程进行编译
bash: make: command not found
1
2

解决方法

yum -y install gcc automake autoconf libtool make
1

安装

make install
1

更新软链接

cd /usr/bin/
mv gcc gcc_back
mv g++ g++_back
ln -s /usr/local/bin/gcc gcc
ln -s /usr/local/bin/g++ g++

1
2
3
4
5
6

f. 推送对应的镜像到仓库

将容器变成镜像

docker commit -a "runoob.com" -m "my apache" a404c6c174a2  mymysql:v1
1

OPTIONS说明：

• -a :提交的镜像作者；

• -c :使用Dockerfile指令来创建镜像；

• -m :提交时的说明文字；

• -p :在commit时，将容器暂停。

将镜像推入仓库

重新定义tag的名字

#docker  tag  镜像id       要推入仓库的用户名/要推入的仓库名:新定义的tag 
docker tag e12a9ec48ab7    felaim/felaim_sever:myproject
1
2

#docker push  要推入仓库的用户名/要推入的仓库名:镜像标签
docker push  felaim/felaim_sever:myproject
1
2

安装gdb

wget ftp.gnu.org/gnu/gdb/gdb-7.8.tar.gz
tar -xzvf gdb-7.8.tar.gz
cd gdb-7.8
./configure
make 
1
2
3
4
5

出现错误

configure: WARNING: no enhanced curses library found; disabling TUI
checking for library containing tgetent... no
configure: error: no termcap library found
make[1]: *** [configure-gdb] Error 1
make[1]: Leaving directory `/usr/local/src/gdb-7.8'
make: *** [all] Error 2

1
2
3
4
5
6
7

yum -y install ncurses-devel
1

重新configu后,make,make install

安装cmake

For example, if you simply want to build and install CMake from source,
you can build directly in the source tree::

  $ ./bootstrap && make && sudo make install

1
2
3
4
5

报错

gmake: `cmake' is up to date.
/usr/local/src/cmake-3.15.7/Bootstrap.cmk/cmake: /lib64/libstdc++.so.6: version `GLIBCXX_3.4.20' not found (required by /usr/local/src/cmake-3.15.7/Bootstrap.cmk/cmake)
/usr/local/src/cmake-3.15.7/Bootstrap.cmk/cmake: /lib64/libstdc++.so.6: version `GLIBCXX_3.4.21' not found (required by /usr/local/src/cmake-3.15.7/Bootstrap.cmk/cmake)
1
2
3

解决方法

cp /usr/local/lib64/libstdc++.so.6.0.21 /usr/lib64/
cd /usr/lib64/
rm -f libstdc++.so.6
ln -s libstdc++.so.6.0.21 libstdc++.so.6   
1
2
3
4

问题解决后运行gmake

gmake
make install
cmake -version
1
2
3

安装locate

bash: locate: command not found
yum install mlocate
1
2

出现问题

locate: can not stat () `/var/lib/mlocate/mlocate.db': No such file or directory
updatedb
1
2

安装protobuf (2020.4.20)

安装tf1.14对应的protobuf需要3.7.1版本的,所以先提前安装好,这样OpenCV链接之间链接3.7.1版本的即可
下载对应版本的protobuf(git clone怎么选择tag和version)

使用–branch指定分支,也得多试几次,开手机热点比较快

git clone --branch v3.7.1 https://github.com/protocolbuffers/protobuf.git

cd protobuf

//配置脚本
./autogen.sh

//编译与安装
./configure

make
make check
sudo make install
sudo ldconfig # refresh shared library cache.
protoc --version
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

出现warning(暂未解决)

=== configuring in third_party/googletest (/home/software/protobuf/third_party/googletest)
configure: WARNING: no configuration information is in third_party/googletest
1
2

安装eigen

下载对应安装包

mkdir build
cd build
cmake ..
make install
1
2
3
4

安装bazel

yum install unzip
./bazel-0.24.1-installer-linux-x86_64.sh --user
export PATH="$PATH:$HOME/bin"
1
2
3

TensorRT

下载对应版本, TensorRT-6.0.1.8.CentOS-7.6.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz

tar -zxvf TensorRT-6.0.1.8.CentOS-7.6.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:
1
2

tensorflow1.14

./configure中path的设置
/usr/local/cuda,/usr/local/cuda/bin,/usr/local/cuda/lib64,/usr/local/cuda/include,/home/software/TensorRT-6.0.1.8,/usr/lib/x86_64-linux-gnu,/usr/include,/usr/lib64

1
2
3

patch command is not found, please install it and referenced by '//third_party/eigen3:eigen3'
yum -y install  patch
1
2

ERROR: /root/.cache/bazel/_bazel_root/79617ab99e45fd3cf89b3b67f497350d/external/nccl_archive/BUILD.bazel:67:1: fatbinary external/nccl_archive/device_dlink_hdrs.fatbin failed (Exit 1)
fatbinary fatal   : Unknown option '-bin2c-path'
Target //tensorflow:libtensorflow_cc.so failed to build
Use --verbose_failures to see the command lines of failed build steps.


1
2
3
4
5
6

3. 安装OpenCV 4.2版本

安装OpenCV4.2版本,安装对应依赖项

sudo yum -y install epel-release
sudo yum -y install qt5-qtbase-devel gtk2-devel libpng-devel jasper-devel openexr-devel libwebp-devel libjpeg-turbo-devel freeglut-devel mesa-libGL mesa-libGL-devel libtiff-devel libdc1394-devel tbb-devel eigen3-devel boost boost-thread boost-devel libv4l-devel gstreamer-plugins-base-devel
1
2

mkdir /home/software/
git clone https://github.com/opencv/opencv.git
git clone https://github.com/opencv/opencv_contrib.git
cp -r ./opencv_contrib ./opencv/
cd opencv
mkdir build
cd build
cmake   -D CMAKE_BUILD_TYPE=RELEASE   -D CMAKE_INSTALL_PREFIX=/usr/local   -D WITH_TBB=ON   -D WITH_V4L=ON   -D WITH_QT=ON   -D WITH_OPENGL=ON   -D WITH_CUDA=ON   -D ENABLE_FAST_MATH=1   -D CUDA_FAST_MATH=1   -D CUDA_NVCC_FLAGS="-D_FORCE_INLINES"   -D WITH_CUBLAS=1 -D OPENCV_EXTRA_MODULES_PATH=../opencv_contrib/modules/ ..

1
2
3
4
5
6
7
8
9

参考地址

gcc, g++升级:

• https://blog.csdn.net/u011181989/article/details/91334478
• https://www.vpser.net/manage/centos-6-upgrade-gcc.html