线上常用shell脚本分享

一.邮件发送

服务器经常可能需要发送邮件，就可以使用mailx工具，通过163、qq等邮件代理发送。
安装

# yum install mailx

设置配置文件

# vi /etc/mail.rc  
set from=qtorm1@163.com smtp=smtp.163.com
set smtp-auth-user=qtorm1@163.com smtp-auth-password=GGAWP*****
set smtp-auth=login

可在163邮箱中找到对应SMTP服务器地址，并设置授权码

二.磁盘监控告警

原理：通关一台控制主机ssh到被监控主机通过df -h查看计算磁盘大小，超过阈值则告警。
首先需要控制主机能免密ssh到被监控主机上，通过ssh-keygen配置

#!/bin/bash

#cat  host.info 
#192.168.1.102  root  22
#192.168.1.103  root  22

HOST_INFO=host.info   #记录ip地址
for IP in $(awk '/^[^#]/{print $1}' $HOST_INFO); do  #除开头#号的所有行第一列
    USER=$(awk -v ip=$IP 'ip==$1{print $2}' $HOST_INFO)
    PORT=$(awk -v ip=$IP 'ip==$1{print $3}' $HOST_INFO)
    TMP_FILE=/tmp/disk.tmp
    ssh -p $PORT $USER@$IP 'df -h' > $TMP_FILE
    USE_RATE_LIST=$(awk 'BEGIN{OFS="="}/^\/dev/{print $NF,int($5)}' $TMP_FILE)
    for USE_RATE in $USE_RATE_LIST; do
        PART_NAME=${USE_RATE%=*}
        USE_RATE=${USE_RATE#*=}
        if [ $USE_RATE -ge 80 ]; then
            echo "Warning: $PART_NAME Partition usage $USE_RATE%!"
        fi
    done
done

三.网卡流量检测

原理：网卡流量可以在系统/proc/net/dev查看

NIC=$1
echo -e " In ------ Out"
while true; do
    OLD_IN=$(awk '$0~"'$NIC'"{print $2}' /proc/net/dev)  #全文匹配($0~)含有$NIC的行,并输出此行的第二列
    OLD_OUT=$(awk '$0~"'$NIC'"{print $10}' /proc/net/dev)
    sleep 1
    NEW_IN=$(awk  '$0~"'$NIC'"{print $2}' /proc/net/dev)
    NEW_OUT=$(awk '$0~"'$NIC'"{print $10}' /proc/net/dev)
    IN=$(printf "%.1f%s" "$((($NEW_IN-$OLD_IN)/1024))" "KB/s")  #%.1f代表保留一位浮点数，%s表示占位符
    OUT=$(printf "%.1f%s" "$((($NEW_OUT-$OLD_OUT)/1024))" "KB/s")
    echo "$IN $OUT"
    sleep 1
done

效果：一秒一次检测某网卡的入和出流量

四.系统初始化优化

常用系统初始化优化如下，可根据自身需求操作

#/bin/bash
#服务器时间同步
ln -s /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
if ! crontab -l |grep ntpdate &>/dev/null ; then
    (echo "* 1 * * * ntpdate time.windows.com >/dev/null 2>&1";crontab -l) |crontab 
fi

# 禁用selinux
sed -i '/SELINUX/{s/permissive/disabled/}' /etc/selinux/config

# 关闭防火墙
if egrep "7.[0-9]" /etc/redhat-release &>/dev/null; then
    systemctl stop firewalld
    systemctl disable firewalld
elif egrep "6.[0-9]" /etc/redhat-release &>/dev/null; then
    service iptables stop
    chkconfig iptables off
fi

# 历史命令显示操作时间
if ! grep HISTTIMEFORMAT /etc/bashrc; then
    echo 'export HISTTIMEFORMAT="%F %T `whoami` "' >> /etc/bashrc
fi

# SSH超时时间设置
if ! grep "TMOUT=600" /etc/profile &>/dev/null; then
    echo "export TMOUT=600" >> /etc/profile
fi

# 禁止root远程登录
sed -i 's/#PermitRootLogin yes/PermitRootLogin no/' /etc/ssh/sshd_config

# 禁止定时任务向发送邮件
sed -i 's/^MAILTO=root/MAILTO=""/' /etc/crontab 

# 设置最大打开文件数
if ! grep "* soft nofile 65535" /etc/security/limits.conf &>/dev/null; then
cat >> /etc/security/limits.conf << EOF
    * soft nofile 65535
    * hard nofile 65535
EOF
fi

# 系统内核优化
cat >> /etc/sysctl.conf << EOF
net.ipv4.tcp_syncookies = 1
net.ipv4.tcp_max_tw_buckets = 20480
net.ipv4.tcp_max_syn_backlog = 20480
net.core.netdev_max_backlog = 262144
net.ipv4.tcp_fin_timeout = 20  
EOF

# 减少SWAP使用 (权重0-100,值越大使用越多)
echo "0" > /proc/sys/vm/swappiness

# 安装系统性能分析工具及其他
yum install gcc make autoconf vim sysstat net-tools iostat iftop iotp lrzsz -y

五.批量创建用户

运行：./脚本名.sh 用户名1 用户名2 …

#!/bin/bash

USER_LIST=$@
USER_FILE=user.txt
for USER in $USER_LIST; do
    if ! id $USER &>/dev/null; then
        PASS=$(echo $RANDOM |md5sum |cut -c 1-8)
        useradd $USER
        echo $PASS |passwd --stdin $USER &>/dev/null
        echo "$USER   $PASS" >> $USER_FILE
        echo "$USER User create successful."
    else
        echo "$USER User already exists!"
    fi
done

六.找cpu内存高的进程

ps -eo可以定义进程项，–sort=是把某一项排序，head显示数量

ps -eo user,pid,pcpu,pmem,args --sort=-pcpu  |head -n 10

ps -eo user,pid,pcpu,pmem,args --sort=-pmem  |head -n 10

七.查看服务器利用率

原理：cpu是通过vmstat命令收集的数据，memory是通过free -m命令收集，磁盘是通过df命令，tcp是通过ss命令。

#!/bin/bash

function cpu() {
    NUM=1
    while [ $NUM -le 3 ]; do
        util=`vmstat |awk '{if(NR==3)print 100-$15"%"}'`
        user=`vmstat |awk '{if(NR==3)print $13"%"}'`
        sys=`vmstat |awk '{if(NR==3)print $14"%"}'`
        iowait=`vmstat |awk '{if(NR==3)print $16"%"}'`
        echo "CPU - 使用率: $util , 等待磁盘IO响应使用率: $iowait"
        let NUM++
        sleep 1
    done
}

function memory() {
    total=`free -m |awk '{if(NR==2)printf "%.1f",$2/1024}'`
    used=`free -m |awk '{if(NR==2) printf "%.1f",($2-$NF)/1024}'`
    available=`free -m |awk '{if(NR==2) printf "%.1f",$NF/1024}'`
    echo "内存 - 总大小: ${total}G , 使用: ${used}G , 剩余: ${available}G"
}

function disk() {
    fs=$(df -h |awk '/^\/dev/{print $1}')
    for p in $fs; do
        mounted=$(df -h |awk '$1=="'$p'"{print $NF}')
        size=$(df -h |awk '$1=="'$p'"{print $2}')
        used=$(df -h |awk '$1=="'$p'"{print $3}')
        used_percent=$(df -h |awk '$1=="'$p'"{print $5}')
        echo "硬盘 - 挂载点: $mounted , 总大小: $size , 使用: $used , 使用率: $used_percent"
    done
}

function tcp_status() {
    summary=$(ss -antp |awk '{status[$1]++}END{for(i in status) printf i":"status[i]" "}')
    echo "TCP连接状态 - $summary"
}

cpu
memory
disk
tcp_status

八.批量检测网站异常

原理：通过curl检测网站，再通过返回的状态码确认是否异常，设定2和3开头的为访问正常。
通过for语句循环判断，超过3次不是上述状态码则可以触发发送邮件等操作。

#!/bin/bash  
URL_LIST="www.baidu.com www.qq.com"
for URL in $URL_LIST; do  #遍历列表网站
    FAIL_COUNT=0
    for ((i=1;i<=3;i++)); do
        HTTP_CODE=$(curl -o /dev/null --connect-timeout 3 -s -w "%{http_code}" $URL)   #输出重定向/dev/null,-w只保留状态码,-s静默模式,超时时间3秒
        if [[ $HTTP_CODE = 2?? ]] || [[ $HTTP_CODE = 3?? ]]; then     #检测状态码2或者3开头的为正常
            echo "$URL OK"
            break
        else
            echo "$URL retry $FAIL_COUNT"
            let FAIL_COUNT++
        fi
    done
    if [ $FAIL_COUNT -eq 3 ]; then   #表示3次访问失败
        echo "Warning: $URL Access failure!"
    fi
done

九.批量主机执行命令

市场上有很多像abible类似的工具,但是这是通过shell的方式ssh到被控制主机执行命令,再通过expect进行交互，有些场景可能会用得到。
# cat host.info
192.168.0.102 root 22 password1
192.168.0.103 root 22 password1

#!/bin/bash

COMMAND=$*
HOST_INFO=host.info
for IP in $(awk '/^[^#]/{print $1}' $HOST_INFO); do
    USER=$(awk -v ip=$IP 'ip==$1{print $2}' $HOST_INFO)
    PORT=$(awk -v ip=$IP 'ip==$1{print $3}' $HOST_INFO)
    PASS=$(awk -v ip=$IP 'ip==$1{print $4}' $HOST_INFO)
    expect -c "
       spawn ssh -p $PORT $USER@$IP
       expect {
          \"(yes/no)\" {send \"yes\r\"; exp_continue}
          \"password:\" {send \"$PASS\r\"; exp_continue}
          \"$USER@*\" {send \"$COMMAND\r exit\r\"; exp_continue}
       }
    "
    echo "-------------------"
done

十.一键部署LNMP

正常选编编译安装nginx、php，可以按需设置安装版本

#!/bin/bash
#按需设置版本
NGINX_V=1.15.6
PHP_V=5.6.36
TMP_DIR=/tmp

INSTALL_DIR=/usr/local

PWD_C=$PWD  #当前路径变量

echo
echo -e "\tMenu\n"
echo -e "1. Install Nginx"
echo -e "2. Install PHP"
echo -e "3. Install MySQL"
echo -e "4. Deploy LNMP"
echo -e "9. Quit"

function command_status_check() {
	if [ $? -ne 0 ]; then
		echo $1
		exit
	fi 
}

function install_nginx() {
    cd $TMP_DIR
    yum install -y gcc gcc-c++ make openssl-devel pcre-devel wget
    wget http://nginx.org/download/nginx-${NGINX_V}.tar.gz
    tar zxf nginx-${NGINX_V}.tar.gz
    cd nginx-${NGINX_V}
    ./configure --prefix=$INSTALL_DIR/nginx \
    --with-http_ssl_module \
    --with-http_stub_status_module \
    --with-stream
    command_status_check "Nginx - 平台环境检查失败！"
    make -j 4 
    command_status_check "Nginx - 编译失败！"
    make install
    command_status_check "Nginx - 安装失败！"
    mkdir -p $INSTALL_DIR/nginx/conf/vhost
    alias cp=cp ; cp -rf $PWD_C/nginx.conf $INSTALL_DIR/nginx/conf
    rm -rf $INSTALL_DIR/nginx/html/*
    echo "ok" > $INSTALL_DIR/nginx/html/status.html
    echo '' > $INSTALL_DIR/nginx/html/status.php
    $INSTALL_DIR/nginx/sbin/nginx
    command_status_check "Nginx - 启动失败！"
}

function install_php() {
	cd $TMP_DIR
    yum install -y gcc gcc-c++ make gd-devel libxml2-devel \
        libcurl-devel libjpeg-devel libpng-devel openssl-devel \
        libmcrypt-devel libxslt-devel libtidy-devel
    wget http://docs.php.net/distributions/php-${PHP_V}.tar.gz
    tar zxf php-${PHP_V}.tar.gz
    cd php-${PHP_V}
    ./configure --prefix=$INSTALL_DIR/php \
    --with-config-file-path=$INSTALL_DIR/php/etc \
    --enable-fpm --enable-opcache \
    --with-mysql --with-mysqli --with-pdo-mysql \
    --with-openssl --with-zlib --with-curl --with-gd \
    --with-jpeg-dir --with-png-dir --with-freetype-dir \
    --enable-mbstring --enable-hash
    command_status_check "PHP - 平台环境检查失败！"
    make -j 4 
    command_status_check "PHP - 编译失败！"
    make install
    command_status_check "PHP - 安装失败！"
    cp php.ini-production $INSTALL_DIR/php/etc/php.ini
    cp sapi/fpm/php-fpm.conf $INSTALL_DIR/php/etc/php-fpm.conf
    cp sapi/fpm/init.d.php-fpm /etc/init.d/php-fpm
    chmod +x /etc/init.d/php-fpm
    /etc/init.d/php-fpm start
    command_status_check "PHP - 启动失败！"
}

read -p "请输入编号：" number
case $number in
    1)
        install_nginx;;
    2)
        install_php;;
    3)
        install_mysql;;
    4)
        install_nginx
        install_php
        ;;
    9)
        exit;;
esac

十一.监控mysql主从同步是否异常

原理：主库写入数据到binlog中，从库会在readylog复制执行到从库中，可以在数据库show slave status;通过监听Slave_IO_Running和Slave_SQL_Running两个状态（Yes）来判断是否异常。

#!/bin/bash  

HOST=localhost
USER=root
PASSWD=123.com
IO_SQL_STATUS=$(mysql -h$HOST -u$USER -p$PASSWD -e 'show slave status\G' 2>/dev/null |awk '/Slave_.*_Running:/{print $1$2}')   #加上2>/dev/null可以去掉warning提示，再通过awk筛选Slave_IO_Running和Slave_SQL_Running两个值
for i in $IO_SQL_STATUS; do     #两个值遍历
    THREAD_STATUS_NAME=${i%:*}  #去掉：后面的得到名称
    THREAD_STATUS=${i#*:}       #去掉：前面的得到状态值
    if [ "$THREAD_STATUS" != "Yes" ]; then   #判断值不等于yes则触发告警发送邮件
        echo "Error: MySQL Master-Slave $THREAD_STATUS_NAME status is $THREAD_STATUS!" |mail -s "Master-Slave Staus" xxx@163.com
    fi
done

十二.mysql分库分表备份

备份命令：
分库：mysqldump -uroot -pxxxx -B A库　> A.sql
分表：mysqldump -uroot -pxxxx A库 table > t.sql
分库备份

#!/bin/bash

DATE=$(date +%F_%H-%M-%S)
HOST=localhost
USER=backup
PASS=123.com
BACKUP_DIR=/data/db_backup
DB_LIST=$(mysql -h$HOST -u$USER -p$PASS -s -e "show databases;" 2>/dev/null |egrep -v "Database|information_schema|mysql|performance_schema|sys")  #egrep去除不要的表，也可以直接指定需要备份的表

for DB in $DB_LIST; do  #遍历备份的库
    BACKUP_NAME=$BACKUP_DIR/${DB}_${DATE}.sql
    if ! mysqldump -h$HOST -u$USER -p$PASS -B $DB > $BACKUP_NAME 2>/dev/null; then
        echo "$BACKUP_NAME 备份失败!"
    fi
done

分表备份

#!/bin/bash

DATE=$(date +%F_%H-%M-%S)
HOST=localhost
USER=backup
PASS=123.com
BACKUP_DIR=/data/db_backup
DB_LIST=$(mysql -h$HOST -u$USER -p$PASS -s -e "show databases;" 2>/dev/null |egrep -v "Database|information_schema|mysql|performance_schema|sys")

for DB in $DB_LIST; do
    BACKUP_DB_DIR=$BACKUP_DIR/${DB}_${DATE}
    [ ! -d $BACKUP_DB_DIR ] && mkdir -p $BACKUP_DB_DIR &>/dev/null
    TABLE_LIST=$(mysql -h$HOST -u$USER -p$PASS -s -e "use $DB;show tables;" 2>/dev/null) #得出要备份的库中所有的表
    for TABLE in $TABLE_LIST; do
        BACKUP_NAME=$BACKUP_DB_DIR/${TABLE}.sql 
        if ! mysqldump -h$HOST -u$USER -p$PASS $DB $TABLE > $BACKUP_NAME 2>/dev/null; then
            echo "$BACKUP_NAME 备份失败!"
        fi
    done
done

十三.切割nginx日志

原理：备份旧的日志文件，再重新生成pid即可生成新的日志文件

#!/bin/bash


LOG_DIR=/www/wwwlogs
YESTERDAY_TIME=$(date -d "yesterday" +%F)
LOG_MONTH_DIR=$LOG_DIR/$(date +"%Y-%m")
LOG_FILE_LIST="www.yj-example.cn.log"
PID="/www/server/nginx/logs/nginx.pid"

for LOG_FILE in $LOG_FILE_LIST; do
    [ ! -d $LOG_MONTH_DIR ] && mkdir -p $LOG_MONTH_DIR
    tar -zcPf $LOG_DIR/${LOG_FILE}.tar.gz  $LOG_DIR/$LOG_FILE
    mv $LOG_DIR/${LOG_FILE}.tar.gz $LOG_MONTH_DIR/${LOG_FILE}_${YESTERDAY_TIME}.tar.gz

done

kill -USR1 $(cat $PID)

十四.DOS攻击防范

原理：通过检测nginx访问日志，当某个IP一分钟访问大于10次时，使用防火墙屏蔽此IP

#!/bin/bash

DATE=$(date +%d/%b/%Y:%H:%M)
LOG_FILE=/usr/local/nginx/logs/demo2.access.log
ABNORMAL_IP=$(tail -n5000 $LOG_FILE |grep $DATE |awk '{a[$1]++}END{for(i in a)if(a[i]>10)print i}')
for IP in $ABNORMAL_IP; do
    if [ $(iptables -vnL |grep -c "$IP") -eq 0 ]; then
        iptables -I INPUT -s $IP -j DROP
        echo "$(date +'%F_%T') $IP" >> /tmp/drop_ip.log
    fi
done

十五.文件更新监控

原理：通过inotifywait命令监控，有创建、更新、修改文件即发送邮件。

#!/bin/bash

MON_DIR=/opt
inotifywait -mqr --format %f -e create $MON_DIR |\
while read files; do
   rsync -avz /opt /tmp/opt
   #echo "$(date +'%F %T') create $files" | mail -s "dir monitor" xxx@163.com
done