Skip to main content

服务器配置及使用

Tiu MoAbout 3 min

配置基础环境（sudo 用户）

基础设置

sudo apt install vim  # 安装vim，vi太难用

sudo apt install net-tools

sudo apt install gpustat

sudo apt install tmux

sudo apt install screen

配置ssh监控端口

$ sudo vim /etc/ssh/sshd_config
'add' ListenAddress 0.0.0.0:5122(YOUR_PORT)

# 例如：
ListenAddress 0.0.0.0:8989
ListenAddress 0.0.0.0:22

$ service sshd restart  # （如果selinux不设置为disabled，则无法生效）

$ sudo netstat -anp|grep sshd  # 查看端口号

# 如果系统升级过ssh，即使修改配置文件/etc/ssh/sshd_config也不会生效，
# 升级openssh后，配置文件被修改到/usr/local/etc/sshd_config，
# 所以修改这个/usr/local/etc/sshd_config

挂载磁盘看情况

# 列出所有磁盘
$ sudo fdisk -l

# 查询UUID
$ sudo blkid /dev/sda3

# 编辑/etc/fstab
$ sudo vim /etc/fstab # 或者用 gedit

# <file system> <mount point> <type> <options> <dump> <pass>
'add' UUID=eb7568d0-fa94-4a1a-9fe6-75ff37faa310 /home/leng/NewDisk ext4 defaults 0 2

# 查看挂载
$ sudo mount -a  # 挂载后原来文件夹中的东西覆盖，取消挂载才能看到
$ df -h # 查看挂载

$ du -sh  # 查看当前文件夹大小

# 移除挂载
$ sudo umount /dev/sda3

# 复制 /home 至 / 
$ mkdir home_bak
$ cp -Rvp /home /home_bak

# 取消/home分区的挂载
$ sudo vim /etc/fstab

'#' UUID=eb7568d0-fa94-4a1a-9fe6-75ff37faa310 /home/leng/NewDisk ext4 defaults 0 2
$ umount /home
$ mv home_bak home
$ umount /home

$ reboot

sudo usermod -d /userHome/guest -m guest

安装 cuda 可选

建议安装 11.7 2023-0506 2132

'https://developer.nvidia.com/cuda-downloads'  # 最新版本

'https://developer.nvidia.com/cuda-toolkit-archive'  # 之前版本

$ wget https://developer.download.nvidia.com/compute/cuda/12.1.0/local_installers/cuda_12.1.0_530.30.02_linux.run
$ sudo sh cuda_12.1.0_530.30.02_linux.run

vim ~/.bashrc

export PATH=/usr/local/cuda-12.1/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-12.1/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

export PATH=/usr/local/cuda/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}


###
export PATH=$PATH:/usr/local/cuda/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64
export LIBRARY_PATH=$LIBRARY_PATH:/usr/local/cuda/lib64

配置conda

下载miniconda

Miniconda — conda documentationopen in new window

终端安装

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
sh Miniconda3-latest-Linux-x86_64.sh
source .bashrc

设置anaconda组

$ sudo groupadd anaconda  # 创建anaconda组
$ sudo adduser 'USERNAME' anaconda # 添加用户进组

# 添加新用户后需再设置一次 权限
$ sudo chgrp -R anaconda /opt # 组内共享这个目录下的文件
$ sudo chmod 777 -R /opt # 设置权限


# 设置数据文件夹
$ sudo chgrp -R anaconda /data
$ sudo chmod 777 -R /data

创建新用户

$ sudo adduser 'USERNAME'
$ sudo usermod -aG 'GROUPNAME' 'USERNAME'

# 删除用户
$ grep bash /etc/passwd # 查看用户
$ sudo deluser 'USERNAME' sudo  # 移除sudo
$ sudo deluser --remove-home 'USERNAME'
sudo groupdel 'GROUPNAME'
$ sudo deluser --remove-all-files 'USERNAME'  # 删除用户所有

.bashrc

sudo cp /home/guest/.bashrc /home/nemo/.bashrc 
sudo chown nemo:nemo /home/nemo/.bashrc

export PATH=$PATH:/usr/local/cuda/bin/

# >>> conda initialize >>>
# !! Contents within this block are managed by 'conda init' !!
__conda_setup="$('/opt/miniconda3/bin/conda' 'shell.bash' 'hook' 2> /dev/null)"
if [ $? -eq 0 ]; then
    eval "$__conda_setup"
else
    if [ -f "/opt/miniconda3/etc/profile.d/conda.sh" ]; then
        . "/opt/miniconda3/etc/profile.d/conda.sh"
    else
        export PATH="/opt/miniconda3/bin:$PATH"
    fi
fi
unset __conda_setup
# <<< conda initialize <<<

.condarc User

vim ~/.condarc
source ~/.bashrc

show_channel_urls: true

channels:
#  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
#  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forge/
#  - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
  - defaults

envs_dirs:
  - /home/'YOURNAME'/.conda/envs
pkgs_dirs:
  - /home/'YOURNAME'/.conda/pkgs

.pip User

$ mkdir .pip && cd .pip && vim pip.conf 
'or' mkdir .pip; cd .pip; vim pip.conf 

[global]
timeout =6000
index-url =http://pypi.douban.com/simple/
[install]
use-mirrors =true 
mirrors =http://pypi.douban.com/simple/ 
trusted-host =pypi.douban.com

conda User

$ conda create -n 'ENV_NMAE'
$ conda activate 'ENV_NMAE'

# https://pytorch.org/get-started/locally/

$ conda install pytorch torchvision torchaudio pytorch-cuda=11.7 -c pytorch -c nvidia
$ pip install opencv-contrib-python

查看cuda及cuDNN能否使用

# 判断是否安装了cuda
$ python
$ import torch
$ print(torch.cuda.is_available())  #返回True则说明已经安装了cuda
# 判断是否安装了cuDNN
$ from torch.backends import  cudnn 
$ print(cudnn.is_available())  #返回True则说明已经安装了cuDNN

ctrl+d 退出

export http_proxy="http://10.16.0.81:8888/"
export https_proxy="http://10.16.0.81:8888/"

export PATH=/usr/local/cuda/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

# >>> conda initialize >>>
# !! Contents within this block are managed by 'conda init' !!
__conda_setup="$('/opt/miniconda3/bin/conda' 'shell.bash' 'hook' 2> /dev/null)"
if [ $? -eq 0 ]; then
    eval "$__conda_setup"
else
    if [ -f "/opt/miniconda3/etc/profile.d/conda.sh" ]; then
        . "/opt/miniconda3/etc/profile.d/conda.sh"
    else
        export PATH="/opt/miniconda3/bin:$PATH"
    fi
fi
unset __conda_setup
# <<< conda initialize <<<
conda activate tai

1.如何把一个挂载好的200g目录，假设是/data 变成/work
（1）df -Th （查看/data挂载目录的磁盘大小有200g满足需求）
得到 /dev/mapper/centos-swap 200.0G 33M 2.000G 2% /data
（2）umount /data （把/data和挂载磁盘卸载）
（3）mv /data /work （对/data文件改名为目标目录/work）
（4）vim /etc/fstab (编辑fstab，实现永久挂载)
UUID=498ccf9b-926b-43f9-9839-2099cd9b0878 /boot xfs defaults 0 0
/dev/mapper/centos-swap /data xfs defaults 0 0
修改：把其中的/data 变成 /work
（5）mount -a

免密链接

# 本地终端
$ ssh-keygen -t rsa -C "ANYTHING"  # 一路 Enter
$ vim "USERPATH/.ssh/id_rsa.pub" #e.g. /Users/tiumo/.ssh/id_rsa.pub # 打开密钥，手动复制

$ cat /Users/tiumo/.ssh/id_rsa.pub | pbcopy  # Mac OS

# 远程终端
$ mkdir .ssh 
$ cd .ssh
$ vim authorized_keys  #/home/leng/.ssh/authorized_keys #手动粘贴
ᕦ(ò_óˇ)ᕤ
Copyright © 2025 Tiu Mo