Shell script is frequently used when monitoring system status on Linux. It's not an easy task to write shell script but it's a very important skill for developers and system administrators to ease work and automate common tasks. This post will share some common interview questions about shell script.

1. Get random characters(8 characters)

Method 1
# echo $RANDOM |md5sum |cut -c 1-8
471b94f2
Method 2
# openssl rand -base64 4
vg3BEg==
Method 3
# cat /proc/sys/kernel/random/uuid |cut -c 1-8
ed9e032c

2. Get random numbers(8 digits)

Method 1
# echo $RANDOM |cksum |cut -c 1-8
23648321
Method 2
# openssl rand -base64 4 |cksum |cut -c 1-8
38571131
Method 3
# date +%N |cut -c 1-8
69024815

3. Define a colored output function

This is useful when need to print logs in different colors which indicate different level of messages(like INFO, WARN etc)

# Method 1
function echo_color() {
    if [ $1 == "green" ]; then
        echo -e "\033[32;40m$2\033[0m"
    elif [ $1 == "red" ]; then
        echo -e "\033[31;40m$2\033[0m"
    fi
}

#Method 2
function echo_color() {
    case $1 in
        green)
            echo -e "\033[32;40m$2\033[0m"
            ;;
        red)
            echo -e "\033[31;40m$2\033[0m" 
            ;;
        *) 
            echo "Example: echo_color red string"
    esac
}

4. Create users in batch

Sometimes system administrators need to create users in batch for some new hires in the company at once. Below can help achieve this.

#!/bin/bash
DATE=$(date +%F_%T)
USER_FILE=user.txt
echo_color(){
    if [ $1 == "green" ]; then
        echo -e "\033[32;40m$2\033[0m"
    elif [ $1 == "red" ]; then
        echo -e "\033[31;40m$2\033[0m"
    fi
}
# Backup the file first if the userfile exists and its size is larger than 0
if [ -s $USER_FILE ]; then
    mv $USER_FILE ${USER_FILE}-${DATE}.bak
    echo_color green "$USER_FILE exist, rename ${USER_FILE}-${DATE}.bak"
fi
echo -e "User\tPassword" >> $USER_FILE
echo "----------------" >> $USER_FILE
for USER in user{1..10}; do
    if ! id $USER &>/dev/null; then
        PASS=$(echo $RANDOM |md5sum |cut -c 1-8)
        useradd $USER
        echo $PASS |passwd --stdin $USER &>/dev/null
        echo -e "$USER\t$PASS" >> $USER_FILE
        echo "$USER User create successful."
    else
        echo_color red "$USER User already exists!"
    fi
done

5. Check whether a package is installed

#!/bin/bash
if rpm -q sysstat &>/dev/null; then
    echo "sysstat is already installed."
else
    echo "sysstat is not installed!"
fi

6. Check service status

#!/bin/bash
PORT_C=$(ss -anu |grep -c 123)
PS_C=$(ps -ef |grep ntpd |grep -vc grep)
if [ $PORT_C -eq 0 -o $PS_C -eq 0 ]; then
    echo "Body" | mail -s "Subject" dst@example.com
fi

7. Monitor CPU, memory and hard drive utilization

a. CPU

Can use vmstat to analyze CPU statistics

#!/bin/bash
DATE=$(date +%F" "%H:%M)
IP=$(ifconfig eth0 |awk -F '[ :]+' '/inet addr/{print $4}')  # Only CentOS6 supports this
MAIL="example@mail.com"
if ! which vmstat &>/dev/null; then
    echo "vmstat command no found, Please install procps package." 
    exit 1
fi
US=$(vmstat |awk 'NR==3{print $13}')
SY=$(vmstat |awk 'NR==3{print $14}')
IDLE=$(vmstat |awk 'NR==3{print $15}')
WAIT=$(vmstat |awk 'NR==3{print $16}')
USE=$(($US+$SY))
if [ $USE -ge 50 ]; then
    echo "
    Date: $DATE
    Host: $IP
    Problem: CPU utilization $USE
    " | mail -s "CPU Monitor" $MAIL
fi

b. Memory

#!/bin/bash
DATE=$(date +%F" "%H:%M)
IP=$(ifconfig eth0 |awk -F '[ :]+' '/inet addr/{print $4}')  
MAIL="example@mail.com"
TOTAL=$(free -m |awk '/Mem/{print $2}')
USE=$(free -m |awk '/Mem/{print $3-$6-$7}')
FREE=$(($TOTAL-$USE))
# 内存小于1G发送报警邮件
if [ $FREE -lt 1024 ]; then
    echo "
    Date: $DATE
    Host: $IP
    Problem: Total=$TOTAL,Use=$USE,Free=$FREE
    " | mail -s "Memory Monitor" $MAIL
fi

c. Hard drive

#!/bin/bash
DATE=$(date +%F" "%H:%M)
IP=$(ifconfig eth0 |awk -F '[ :]+' '/inet addr/{print $4}')  
MAIL="example@mail.com"
TOTAL=$(fdisk -l |awk -F'[: ]+' 'BEGIN{OFS="="}/^Disk \/dev/{printf "%s=%sG,",$2,$3}')
PART_USE=$(df -h |awk 'BEGIN{OFS="="}/^\/dev/{print $1,int($5),$6}')
for i in $PART_USE; do
    PART=$(echo $i |cut -d"=" -f1)
    USE=$(echo $i |cut -d"=" -f2)
    MOUNT=$(echo $i |cut -d"=" -f3)
    if [ $USE -gt 80 ]; then
        echo "
        Date: $DATE
        Host: $IP
        Total: $TOTAL
        Problem: $PART=$USE($MOUNT)
        " | mail -s "Disk Monitor" $MAIL
    fi
done

8. Monitor hard drive utilization in batch

#!/bin/bash
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)
    TMP_FILE=/tmp/disk.tmp
    ssh -p $PORT $USER@$IP 'df -h' > $TMP_FILE
    USE_RATE_LIST=$(awk 'BEGIN{OFS="="}/^\/dev/{print $1,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

9. Check website availability

# Method 1
check_url() {
    HTTP_CODE=$(curl -o /dev/null --connect-timeout 3 -s -w "%{http_code}" $1)
    if [ $HTTP_CODE -ne 200 ]; then
        echo "Warning: $1 Access failure!"
    fi
}

# Method 2
check_url() {
if ! wget -T 10 --tries=1 --spider $1 >/dev/null 2>&1; then  
        echo "Warning: $1 Access failure!"
    fi
}

10. Check MySQL master/slave sync status

#!/bin/bash  
USER=bak
PASSWD=123456
IO_SQL_STATUS=$(mysql -u$USER -p$PASSWD -e 'show slave status\G' |awk -F: '/Slave_.*_Running/{gsub(": ",":");print $0}')  #gsub去除冒号后面的空格
for i in $IO_SQL_STATUS; do
    THREAD_STATUS_NAME=${i%:*}
    THREAD_STATUS=${i#*:}
    if [ "$THREAD_STATUS" != "Yes" ]; then
        echo "Error: MySQL Master-Slave $THREAD_STATUS_NAME status is $THREAD_STATUS!"
    fi
done

These scripts are useful in real work for system administrators. The questions behind these scripts are also frequently asked during interviews.

Reference

http://blog.51cto.com/lizhenliang/1929044