# 1.BIOS设置 #
##############
# 登录iBMC网站进入服务器
# 在登录服务器时进入BIOS
##############
# 2.系统重装 #
##############
# 进入虚拟化界面，载入Ubuntu镜像
# 重装Ubuntu系统，设置root密码
#############
# 3.配置网卡 #
#############
ip addr #查看网卡状态
vi /etc/sysconfig/network-scripts/ifcfg-eno1 #修改网卡配置信息
service network restart #重启网络服务 systemctl restart network
systemctl status network #查看网络状态
#如果网路重启失败，请尝试以下命令
systemctl stop NetworkManager
systemctl disable NetworkManager
###########
# 4.配置代理 #
##############
vi /etc/profile
export http_proxy=http://z00575241:zch19950329%2a@90.253.64.225:6688
export https_proxy=https://z00575241:zch19950329%2a@90.253.64.225:6688
export ftp_proxy=ftp://z00575241:zch19950329%2a@90.253.64.225:6688
source /etc/profile
vi /etc/profile
export http_proxy=http://工号:密码@本地IP:6688
export https_proxy=https://工号:密码@本地IP:6688
export ftp_proxy=ftp://工号:密码@本地IP:6688
source /etc/profile
#############
# 5.配置DNS #
#############
vi /etc/resolv.conf
nameserver 8.8.8.8 #Google
nameserver 114.114.114.114 #中国电信
nameserver 223.5.5.5 #阿里
nameserver 2223.6.6.6 #阿里
###############
# 6.更换镜像源 #
###############
#6.1 添加华为的镜像源
mkdir /root/.pip
echo "[global]" > /root/.pip/pip.conf
echo "index-url=https://repo.huaweicloud.com/repository/pypi/simple" >> /root/.pip/pip.conf
echo "trusted-host=repo.huaweicloud.com" >> /root/.pip/pip.conf
echo "timeout=120" >> /root/.pip/pip.conf
cat /root/.pip/pip.conf
#6.2 添加阿里的镜像源
mv /etc/yum.repos.d/CentOS-Base.repo /etc/yum.repos.d/CentOS-Base.repo.backup #进行备份
curl -o /etc/yum.repos.d/CentOS-Base.repo https://mirrors.aliyun.com/repo/Centos-7.repo #下载镜像
yum clean all #清除缓存
yum makecache #生成缓存
###################
# 7.安装图像化桌面 #
###################
yum groupinstall "GNOME Desktop" "Graphical Administration Tools" #安装Gnome包
systemctl get-default #获取当前系统运行形式，会显示multi-user.target（命令行终端）
systemctl set-default graphical.target #设置默认启动为图形界面，reboot后界面会自动是图形窗口
systemctl set-default multi-user.target #换回命令界面启动
reboot
##########################
# 8.安装CANN软件-运行环境 #
##########################
#8.1 检查root用户的umask值
umask
#8.2 创建HwHiAiUser用户
#增加用户
groupadd HwHiAiUser ##创建HwHiAiUser用户属组
useradd -g HwHiAiUser -m -d /home/HwHiAiUser HwHiAiUser ##创建HwHiAiUser用户，其属组为HwHiAiUser
passwd HwHiAiUser #修改HwHiAiUser用户的密码
#修改UID为1000
usermod -u 1002 joey
groupmod -g 1002 joey
usermod -u 1000 HwHiAiUser
groupmod -g 1000 HwHiAiUser
#增加HwHiAiUser 后，home没有HwHiAiUser目录，执行以下命令
cp -a /etc/skel /home/HwHiAiUser
chown -R HwHiAiUser:HwHiAiUser
#8.3 配置源
rm -f /var/run/yum.pid #关闭yum进程
yum makecache #生成缓存
#8.4 配置最大线程数
vi /etc/profile
ulimit -u unlimited
source /etc/profile
#8.5 安装依赖
#8.5.1 检查系统是否安装python依赖以及gcc等软件
#安装依赖
yum install -y gcc make cmake unzip zlib-devel libffi-devel openssl-devel pciutils net-tools sqlite-devel blas-devel lapack-devel openblas-devel gcc-gfortran
#检查依赖
gcc --version
make --version
cmake --version
rpm -qa |grep unzip
rpm -qa |grep zlib-devel
rpm -qa |grep libffi-devel
rpm -qa |grep openssl-devel
rpm -qa |grep pciutils
rpm -qa |grep net-tools
rpm -qa |grep sqlite-devel
rpm -qa |grep blas-devel
rpm -qa |grep lapack-devel
rpm -qa |grep openblas-devel
rpm -qa |grep gcc-gfortran
#正确输出
gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-39)
GNU Make 3.82
cmake version 2.8.12.2
unzip-6.0-21.el7.x86_64
zlib-devel-1.2.7-18.el7.x86_64
libffi-devel-3.0.13-18.el7.x86_64
openssl-devel-1.0.2k-19.el7.x86_64
pciutils-3.5.1-3.el7.x86_64
net-tools-2.0-0.25.20131004git.el7.x86_64
sqlite-devel-3.7.17-8.el7_7.1.x86_64
blas-devel-3.4.2-8.el7.x86_64
lapack-devel-3.4.2-8.el7.x86_64
openblas-devel-0.3.3-2.el7.x86_64
gcc-gfortran-4.8.5-39.el7.x86_64
#8.5.2 安装cmake3.5.2
yum -y install gcc gcc-c++
wget https://cmake.org/files/v3.5/cmake-3.5.2.tar.gz --no-check-certificate
tar -zxvf cmake-3.5.2.tar.gz
cd cmake-3.5.2
./bootstrap --prefix=/usr
make
make install
cmake --version
#8.5.3 安装gcc7.3.0
yum -y install centos-release-scl
yum -y install devtoolset-7-gcc devtoolset-7-gcc-c++ devtoolset-7-binutils
scl enable devtoolset-7 bash
#8.5.4 检查系统是否安装python开发环境
wget https://www.python.org/ftp/python/3.7.5/Python-3.7.5.tgz #使用wget下载python3.7.5源码包
tar -zxvf Python-3.7.5.tgz #解压源码包
cd Python-3.7.5
./configure --prefix=/usr/local/python3.7.5 --enable-shared #配置
make #编译
make install #安装
cp /usr/local/python3.7.5/lib/libpython3.7m.so.1.0 /usr/lib64
#设置软链接
ln -s /usr/local/python3.7.5/bin/python3 /usr/bin/python3.7
ln -s /usr/local/python3.7.5/bin/pip3 /usr/bin/pip3.7
ln -s /usr/local/python3.7.5/bin/python3 /usr/bin/python3.7.5
ln -s /usr/local/python3.7.5/bin/pip3 /usr/bin/pip3.7.5
#查看安装版本
python3.7.5 --version
pip3.7.5 --version
#8.5.5 安装相关依赖
export CFLAGS=-std=c99
pip3.7 install numpy==1.17.2
pip3.7 install decorator
pip3.7 install sympy==1.4
pip3.7 install cffi==1.12.3
pip3.7 install pyyaml
pip3.7 install pathlib2
pip3.7 install grpcio
pip3.7 install grpcio-tools
pip3.7 install protobuf
pip3.7 install scipy
pip3.7 install requests
#8.6 安装昇腾芯片驱动固件和训练软件（易安装工具）
tar -xzvf ascend-sdk-manager-20.0.0.SPC100-x86-centos.tgz
/home/HwHiAiUser/ascend-sdk-manager app install /home/install train --mode=auto --silent=y #successfully installed packages:['nnae', 'plugin', 'toolbox']
#8.7 安装昇腾芯片驱动固件和训练软件（命令行方式）
#安装dkms工具-驱动源码自动编译
yum install gcc
yum install kernel-headers-$(uname -r)
yum install kernel-devel-$(uname -r)
yum install dkms
#增加安装用户对软件包的可执行权限
chmod +x ./A800-9010-NPU_Driver-20.0.RC1.SPC100-X86_64-CentOS7.6_gcc7.3.0.run
chmod +x ./A800-9010-NPU_Firmware-1.73.1105.1.b050.run
#校验软件包安装文件的一致性和完整性
./A800-9010-NPU_Driver-20.0.RC1.SPC100-X86_64-CentOS7.6_gcc7.3.0.run --check
./A800-9010-NPU_Firmware-1.73.1105.1.b050.run --check
#安装驱动和固件
./A800-9010-NPU_Driver-20.0.RC1.SPC100-X86_64-CentOS7.6_gcc7.3.0.run --run
./A800-9010-NPU_Firmware-1.73.1105.1.b050.run --run
#重启运行环境
reboot
#查看安装的驱动版本号
cd /usr/local/Ascend/driver
cat version.info
cd /usr/local/Ascend/firmware
cat version.info
/usr/local/Ascend/driver/tools/upgrade-tool --device_index -1 --component -1 --version #查看安装的NPU固件版本号
npu-smi info #查看NPU工具安装是否成功
#安装Tensorflow 1.15.0
pip3.7 install Tensorflow-cpu==1.15
#安装训练软件
./Ascend-NNAE-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --check
./Ascend-TFPlugin-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --check
./Ascend-Toolbox-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --check
./Ascend-NNAE-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --install
./Ascend-TFPlugin-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --install
./Ascend-Toolbox-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --install
#8.9 安装后处理
#8.9.1 安装haveged 服务
rpm -ivh haveged-1.9.13-1.el7.x86_64.rpm
systemctl enable haveged
systemctl start haveged
#8.9.2 安装python版本的proto
#检查系统中是否存在“/usr/local/python3.7.5/lib/python3.7/site-packages/google/protobuf/pyext/_message.cpython-37m-<arch>-linux-gnu.so这个动态库，若没有则执行以下命令
pip3.7 uninstall protobuf
wget https://github.com/protocolbuffers/protobuf/releases/ #下载3.11.3版本protobuf-python-3.11.3.tar.gz
#以root用户安装protobuf
cd protobuf
yum install autoconf automake libtool curl make gcc-c++ unzip libffi-devel -y
chmod +x autogen.sh
./autogen.sh
./configure --prefix=/usr/local/protobuf
make
make install
ldconfig #刷新共享库
#检查是否安装完成
ln -s /protobuf/bin/protoc /usr/bin/protoc
protoc --version
#安装protobuf的python版本运行库
python3.7 setup.py build --cpp_implementation #进入protobuf软件包目录的python子目录，编译python版本的运行库
cd .. && make install #安装动态库
python3.7 setup.py install --cpp_implementation #进入python子目录，安装python版本的运行库
#检查系统中是否存在“/usr/local/python3.7.5/lib/python3.7/site-packages/protobuf-3.11.3-py3.7-linux-aarch64.egg/google/protobuf/pyext/_message.cpython-37m-<arch>-linux-gnu.so”这个动态库
export LD_PATH_LIBRARY=/protobuf/lib #在运行脚本中增加环境变量的设置
ln -s /protobuf/lib/libprotobuf.so.22.0.3 /usr/lib/libprotobuf.so.22 #建立软连接
#8.9.3 配置device 的网卡IP
pip3.7 install hccn_tool
hccn_tool -i 0 -ip -s address 192.168.100.2 netmask 255.255.255.0
hccn_tool -i 1 -ip -s address 192.168.101.2 netmask 255.255.255.0
hccn_tool -i 2 -ip -s address 192.168.102.2 netmask 255.255.255.0
hccn_tool -i 3 -ip -s address 192.168.103.2 netmask 255.255.255.0
hccn_tool -i 4 -ip -s address 192.168.100.3 netmask 255.255.255.0
hccn_tool -i 5 -ip -s address 192.168.101.3 netmask 255.255.255.0
hccn_tool -i 6 -ip -s address 192.168.102.3 netmask 255.255.255.0
hccn_tool -i 7 -ip -s address 192.168.103.3 netmask 255.255.255.0
#8.9.4 安装后检查
#添加环境变量
vi ~/.bashrc
export toolbox_install_path=/usr/local/Ascend/toolbox/latest/x86_64-linux_gcc7.3.0
export nnae_install_path=/usr/local/Ascend/nnae/latest/x86_64-linux_gcc7.3.0
export tfplugin_install_path=/usr/local/Ascend/tfplugin/latest/x86_64-linux_gcc7.3.0
export toolkit_install_path=/usr/local/Ascend/ascend-toolkit/latest/x86_64-linux_gcc7.3.0
export PATH=/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/sbin:/usr/local/protobuf/bin/:/usr/local/gcc7.3.0/:/usr/local/gcc7.3.0/lib64:${toolbox_install_path}/dmi/bin:${nnae_install_path}/fwkacllib/ccec_compiler/bin:${toolkit_install_path}/atc/bin:${toolkit_install_path}/atc/ccec_compiler/bin
export PYTHONPATH=/home/zch/Scripts/SSD/ModelZoo_SSD-Resnet34_TF_Atlas/00-access:${tfplugin_install_path}/tfplugin/python/site-packages:${tfplugin_install_path}/tfplugin/python/site-packages/npu_bridge:${nnae_install_path}/fwkacllib/python/site-packages:${nnae_install_path}/fwkacllib/python/site-packages/te:${nnae_install_path}/fwkacllib/python/site-packages/topi:${nnae_install_path}/fwkacllib/python/site-packages/hccl:${nnae_install_path}/opp/op_impl/built-in/ai_core/tbe
export LD_LIBRARY_PATH=/usr/lib:/usr/local:/usr/local/lib:/usr/local/dcmi:/usr/local/Ascend/add-ons:${toolbox_install_path}/dmi/lib64:${nnae_install_path}/fwkacllib/lib64:/usr/local/Ascend/driver/lib64/common:/usr/local/Ascend/driver/lib64/driver
export ASCEND_OPP_PATH=${nnae_install_path}/opp
source ~/.bashrc
#检查设备健康状态
ascend-dmi info
ascend-dmi -dg -c {card-number} -l 1
#检查环境软硬件兼容性
ascend-dmi -c
##########################
# 9.安装CANN软件-开发环境 #
##########################
#9.1 安装开发套件（命令行方式）
chmod +x Ascend-Toolkit-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run
./Ascend-Toolkit-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --check
./Ascend-Toolkit-20.0.0.SPC100-x86_64-linux_gcc7.3.0.run --install
