我們在運維中,尤其是linux運維,都知道腳本的重要性歌豺,腳本會讓我們的 運維事半功倍,所以學(xué)會寫腳本是我們每個linux運維必須學(xué)會的一門功課心包,這里收藏linux運維常用的腳本世曾。如何學(xué)好腳本,最關(guān)鍵的是就是大量的練習(xí) 和實踐。根據(jù)以下腳本我們可以拓展轮听,這樣我們提高的很快!舉一反三岭佳!
1.用Shell編程血巍,判斷一文件是不是字符設(shè)備文件,如果是將其拷貝到 /dev 目錄下珊随。
#!/bin/sh
FILENAME=
echo “Input file name:”
read FILENAME
if [ -c "$FILENAME" ]
then
cp $FILENAME /dev
fi
2.設(shè)計一個shell程序述寡,添加一個新組為class1,然后添加屬于這個組的30個用戶叶洞,用戶名的形式為stdxx鲫凶,其中xx從01到30。
#!/bin/sh
i=1
groupadd class1
while [ $i -le 30 ]
do
if [ $i -le 9 ] ;then
USERNAME=stu0${i}
else
USERNAME=stu${i}
fi
useradd $USERNAME
mkdir /home/$USERNAME
chown -R $USERNAME /home/$USERNAME
chgrp -R class1 /home/$USERNAME
i=$(($i+1))
done
3.編寫shell程序衩辟,實現(xiàn)自動刪除50個賬號的功能螟炫。賬號名為stud1至stud50。
#!/bin/sh
i=1
while [ $i -le 50 ]
do
userdel -r stud${i}
i=$(($i+1 ))
done
4.某系統(tǒng)管理員需每天做一定的重復(fù)工作艺晴,請按照下列要求昼钻,編制一個解決方案:
(1)在下午4 :50刪除/abc目錄下的全部子目錄和全部文件;
(2)從早8:00~下午6:00每小時讀取/xyz目錄下x1文件中每行第一個域的全部數(shù)據(jù)加入到/backup目錄下的bak01.txt文件內(nèi)封寞;
(3)每逢星期一下午5:50將/data目錄下的所有目錄和文件歸檔并壓縮為文件:backup.tar.gz然评;
(4)在下午5:55將IDE接口的CD-ROM卸載(假設(shè):CD-ROM的設(shè)備名為hdc);
(5)在早晨8:00前開機后啟動狈究。
參考答案:
解決方案:
(1)用vi創(chuàng)建編輯一個名為prgx的crontab文件碗淌;
(2)prgx文件的內(nèi)容:
50 16 * * * rm -r /abc/*
0 8-18/1 * * * cut -f1 /xyz/x1 >;>; /backup/bak01.txt
50 17 * * * tar zcvf backup.tar.gz /data
55 17 * * * umount /dev/hdc
(3)由超級用戶登錄,用crontab執(zhí)行 prgx文件中的內(nèi)容:
root@xxx:#crontab prgx抖锥;在每日早晨8:00之前開機后即可自動啟動crontab亿眠。
5.設(shè)計一個shell程序,在每月第一天備份并壓縮/etc目錄的所有內(nèi)容宁改,存放在/root/bak目錄里缕探,且文件名為如下形式y(tǒng)ymmdd_etc,yy為年还蹲,mm為月爹耗,dd為日。Shell程序fileback存放在/usr/bin目錄下谜喊。
參考答案:
(1)編寫shell程序fileback:
DIRNAME=`ls /root | grep bak`
if [ -z "$DIRNAME" ] ; then
mkdir /root/bak
cd /root/bak
fi
YY=`date +%y`
MM=`date +%m`
DD=`date +%d`
BACKETC=$YY$MM$DD_etc.tar.gz
tar zcvf $BACKETC /etc
echo “fileback finished!”
(2)編寫任務(wù)定時器:
echo “0 0 1 * * /bin/sh /usr/bin/fileback” >; /root/etcbakcron
crontab /root/etcbakcron
或使用crontab -e 命令添加定時任務(wù):
0 1 * * * /bin/sh /usr/bin/fileback
6.有一普通用戶想在每周日凌晨零點零分定期備份/user/backup到/tmp目錄下潭兽,該用戶應(yīng)如何做?
參考答案:
(1)第一種方法:
用戶應(yīng)使用crontab –e 命令創(chuàng)建crontab文件斗遏。格式如下:
0 0 * * sun cp –r /user/backup /tmp
(2)第二種方法:
用戶先在自己目錄下新建文件file山卦,文件內(nèi)容如下:
0 * * sun cp –r /user/backup /tmp
然后執(zhí)行 crontab file 使生效。
7.設(shè)計一個Shell程序诵次,在/userdata目錄下建立50個目錄账蓉,即user1~user50枚碗,并設(shè)置每個目錄的權(quán)限,其中其他用戶的權(quán)限為:讀铸本;文件所有者的權(quán)限為:讀肮雨、寫、執(zhí)行箱玷;文件所有者所在組的權(quán)限為:讀怨规、執(zhí)行。
參考答案: 建立程序 Pro16如下:
#!/bin/sh
i=1
while [ i -le 50 ]
do
if [ -d /userdata ];then
mkdir -p /userdata/user$i
chmod 754 /userdata/user$i
echo “user$i”
let “i = i + 1″ (或i=$(($i+1))
else
mkdir /userdata
mkdir -p /userdata/user$i
chmod 754 /userdata/user$i
echo “user$i”
let “i = i + 1″ (或i=$(($i+1))
fi
done
8锡足、mysql備份實例波丰,自動備份mysql,并刪除30天前的備份文件
#!/bin/sh
#auto backup mysql
#wugk 2012-07-14
#PATH DEFINE
BAKDIR=/data/backup/mysql/`date +%Y-%m-%d`
MYSQLDB=www
MYSQLPW=backup
MYSQLUSR=backup
if[ $UID -ne 0 ];then
echo This script must use administrator or root user ,please exit!
sleep 2
exit 0
fi
if[ ! -d $BAKDIR ];then
mkdir -p $BAKDIR
else
echo This is $BAKDIR exists ,please exit ….
sleep 2
exit
fi
###mysqldump backup mysql
/usr/bin/mysqldump -u$MYSQLUSR -p$MYSQLPW -d $MYSQLDB >/data/backup/mysql/`date +%Y-%m-%d`/www_db.sql
cd $BAKDIR ; tar -czf www_mysql_db.tar.gz *.sql
cd $BAKDIR ;find . -name “*.sql” |xargs rm -rf[ $? -eq 0 ]&&echo “This `date +%Y-%m-%d` RESIN BACKUP is SUCCESS”
cd /data/backup/mysql/ ;find . -mtime +30 |xargs rm -rf
9舶得、自動安裝Nginx腳本掰烟,采用case方式,選擇方式扩灯,也可以根據(jù)實際需求改成自己想要的腳本
#!/bin/sh
###nginx install shell
###wugk 2012-07-14
###PATH DEFINE
SOFT_PATH=/data/soft/
NGINX_FILE=nginx-1.2.0.tar.gz
DOWN_PATH=http://nginx.org/download/
if[ $UID -ne 0 ];then
echo This script must use administrator or root user ,please exit!
sleep 2
exit 0
fi
if[ ! -d $SOFT_PATH ];then
mkdir -p $SOFT_PATH
fi
download ()
{
cd $SOFT_PATH ;wget $DOWN_PATH/$NGINX_FILE
}
install ()
{
yum install pcre-devel -y
cd $SOFT_PATH ;tar xzf $NGINX_FILE ;cd nginx-1.2.0/ &&./configure –prefix=/usr/local/nginx/ –with-http_stub_status_module –with-http_ssl_module
[ $? -eq 0 ]&&make &&make install
}
start ()
{
lsof -i :80[ $? -ne 0 ]&&/usr/local/nginx/sbin/nginx
}
stop ()
{
ps -ef |grep nginx |grep -v grep |awk ‘{print $2}’|xargs kill -9
}
exit ()
{
echo $? ;exit
}
###case menu #####
case $1 in
download )
download
;;
install )
install
;;
start )
start
;;
stop )
stop
;;
* )
echo “USAGE:$0 {download or install or start or stop}”
exit
esac
10媚赖、批量解壓tar腳本,批量解壓zip并且建立當(dāng)前目錄珠插。
#!/bin/sh
PATH1=/tmp/images
PATH2=/usr/www/images
for i in `ls ${PATH1}/*`
do
tar xvf $i -C $PATH2
done
這個腳本是針對所有tar文件在一個目錄惧磺,但是實際情況中,有可能在下級或者更深的目錄捻撑,我們可以使用find查找
#!/bin/sh
PATH1=/tmp/images
PATH2=/usr/www/images
for i in `find $PATH1 -name ”*.tar” `
do
tar xvf $i -C $PATH2
done
如何是zip文件磨隘,例如123189.zip 132342.zip 等等批量文件,默認(rèn)unzip直接解壓不帶自身目錄顾患,意思是解壓123189.zip完當(dāng)前目錄就是圖片番捂,不能創(chuàng)建123189目錄下并解壓,可以用shell腳本實現(xiàn)
#!/bin/sh
PATH1=/tmp/images
PATH2=/usr/www/images
cd $PATH1
for i in `find . -name ”*.zip”|awk -F. {print $2} `
do
mkdir -p PATH2$i
unzip -o .$i.zip -d PATH2$i
done
ssh 批量上傳文件
上傳文件大多數(shù)用的是ftp江解,但是用ftp有一點不好设预,就是本地和遠(yuǎn)程的目錄要對應(yīng),這樣就要在多個目錄下去切換犁河,這樣挺麻煩的鳖枕,如果不注意的話,很有可能傳錯桨螺。所以想了個辦法利用scp來批量上傳文件或者目錄宾符。
一,scp上傳不要輸入密碼
如果要用scp來上傳文件灭翔,第一步就要去掉scp上傳時要輸入密碼魏烫。要不然就沒辦法批量上傳了。具體請參考:ssh 不用輸入密碼
二,ssh批量上傳腳本
1,要上傳的文件列表放到一個test文件中
root@ubuntu:/home/zhangy# cat test
/home/zhangy/test/aaa
/home/zhangy/test/nginx.conf
/home/zhangy/test/test.sql
/home/zhangy/test/pa.txt
/home/zhangy/test/password
上面就要上傳的文件哄褒。
2,批量上傳的腳本
vim file_upload.sh
#!/bin/sh
DATE=`date +%Y_%m_%d_%H`
if [ $1 ]
then
for file in $(sed '/^$/d' $1) //去掉空行
do
if [ -f $file ] //普通文件
then
res=`scp $file $2:$file` //上傳文件
if [ -z $res ] //上傳成功
then
echo $file >> ${DATE}_upload.log //上傳成功的日志
fi
elif [ -d $file ] //目錄
then
res=`scp -r $file $2:$file`
if [ -z $res ]
then
echo $file >> ${DATE}_upload.log
fi
fi
done
else
echo "no file" >> ${DATE}_error.log
fi
上傳成功后稀蟋,返回的是一個空行,上傳不成功呐赡,什么都不返回
3糊治,上傳的格式
./file_upload.sh test 192.168.1.13
test是上傳列表文件,192.168.1.13文件要傳到的地方罚舱。
++++++++++++++++++++++++++++++++++++++++++++++++++++
1. 轉(zhuǎn)換文件大小寫:
A script to convert the specified filenames to lower case.
#!/bin/sh
# lowerit
# convert all file names in the current directory to lower case
# only operates on plain files--does not change the name of directories
# will ask for verification before overwriting an existing file
for x in `ls`
do
if [ ! -f $x ]; then
continue
fi
lc=`echo $x | tr '[A-Z]' '[a-z]'`
if [ $lc != $x ]; then
mv -i $x $lc
fi
done
or
if test $# = 0
then
echo "Usage $0: <files>" 1>&2
exit 1
fi
for filename in "$@"
do
new_filename=`echo "$filename" | tr A-Z a-z`
test "$filename" = "$new_filename" && continue
if test -r "$new_filename"
then
echo "$0: $new_filename exists" 1>&2
elif test -e "$filename"
then
mv "$filename" "$new_filename"
else
echo "$0: $filename not found" 1>&2
fi
done
2. 看網(wǎng)站 Watch a Website
A script to repeated download a webpage until it matches a regex then notify an e-mail address.
For example to get e-mail when Kesha tickets (not for yourself of course) go on sale you might run:
% watch_website.sh http://ticketek.com.au/ 'Ke[sS$]+ha' andrewt@cse.unsw.edu.au
repeat_seconds=300 #check every 5 minutes
if test $# = 3
then
url=$1
regexp=$2
email_address=$3
else
echo "Usage: $0 <url> <regex>" 1>&2
exit 1
fi
while true
do
if wget -O- -q "$url"|egrep "$regexp" >/dev/null
then
echo "Generated by $0" | mail -s "$url now matches $regexp" $email_address
exit 0
fi
sleep $repeat_seconds
done
3. 轉(zhuǎn)GIF到PNG convert GIF files to PNG
This scripts converts GIF files to PNG files via the intermediate PPM format.
if [ $# -eq 0 ]
then
echo "Usage: $0 files..." 1>&2
exit 1
fi
if ! type giftopnm 2>/dev/null
then
echo "$0: conversion tool giftopnm not found " 1>&2
exit 1
fi
# missing "in ..." defaults to in "$@"
for f
do
case "$f" in
*.gif)
# OK, do nothing
;;
*)
echo "gif2png: skipping $f, not GIF"
continue
;;
esac
dir=`dirname "$f"`
base=`basename "$f" .gif`
result="$dir/$base.png"
giftopnm "$f" | pnmtopng > $result && echo "wrote $result"
done
4. 計數(shù) Counting
A utility script to print the sub-range of integers specified by its arguments.
Useful to use on the command line or from other scripts
if test $# = 1
then
start=1
finish=$1
elif test $# = 2
then
start=$1
finish=$2
else
echo "Usage: $0 <start> <finish>" 1>&2
exit 1
fi
for argument in "$@"
do
if echo "$argument"|egrep -v '^-?[0-9]+$' >/dev/null
then
echo "$0: argument '$argument' is not an integer" 1>&2
exit 1
fi
done
number=$start
while test $number -le $finish
do
echo $number
number=`expr $number + 1` # or number=$(($number + 1))
done
5. 字頻率 Word Frequency
Count the number of time each different word occurs in the files given as arguments.
sed 's/ /\n/g' "$@"| # convert to one word per line
tr A-Z a-z| # map uppercase to lower case
sed "s/[^a-z']//g"| # remove all characters except a-z and '
egrep -v '^$'| # remove empty lines
sort| # place words in alphabetical order
uniq -c| # use uniq to count how many times each word occurs
sort -n # order words in frequency of occurrance
For example
% cd /home/cs2041/public_html/lec/shell/examples
% ./word_frequency.sh dracula.txt|tail
2124 it
2440 that
2486 in
2549 he
2911 a
3600 of
4448 to
4740 i
5833 and
7843 the
6. Finding
Search $PATH for the specified programs
if test $# = 0
then
echo "Usage $0: <program>" 1>&2
exit 1
fi
for program in "$@"
do
program_found=''
for directory in `echo "$PATH" | tr ':' ' '`
do
f="$directory/$program"
if test -x "$f"
then
ls -ld "$f"
program_found=1
fi
done
if test -z $program_found
then
echo "$program not found"
fi
done
Alternative implementation using while, and a cute use of grep and ||
if test $# = 0
then
echo "Usage $0: <program>" 1>&2
exit 1
fi
for program in "$@"
do
echo "$PATH"|
tr ':' '\n'|
while read directory
do
f="$directory/$program"
if test -x "$f"
then
ls -ld "$f"
fi
done|
egrep '.' || echo "$program not found"
done
And another implementation using while, and a cute use of grep and ||
if test $# = 0
then
echo "Usage $0: <program>" 1>&2
exit 1
fi
for program in "$@"
do
n_path_components=`echo $PATH|tr -d -c :|wc -c`
index=1
while test $index -le $n_path_components
do
directory=`echo "$PATH"|cut -d: -f$index`
f="$directory/$program"
if test -x "$f"
then
ls -ld "$f"
program_found=1
fi
index=`expr $index + 1`
done
test -n $program_found || echo "$program not found"
done