目录
修改配置项的值
multi_cluster_set_ha = yes 改为multi_cluster_set_ha = no
命令:
方法1:(这个需要参数名前面不能有空格)
sed -i "s/^参数名=.*/参数名=新值/" 配置文件路径
方法2:
把namelist.txt中user_name:后面的改成:Liudehua (注意:\1表示替换)
sed -r -i "s/(user_name:).*/\1Liudehua/" namelist.txt
把namelist.txt中user_name=后面的改成:Liudehua (注意:\1表示替换)
sed -r -i "s/(user_name=).*/\1Liudehua/" namelist.txt
从上面的例子可以知道(user_name=)表示匹配user_name=,然后.*表示任意个字符,
\1表示替换,连起来就是把匹配到的“user_name=”后面的任意个字符串替换成“Liudehua”
关于sed的sed元字符集的含义更多的见:https://www.cnblogs.com/maxincai/p/5146338.html
例子一:
要将vim /etc/onestor/onestor.conf中的
multi_cluster_set_ha = yes
handy_ha_needed = yes
改为:
multi_cluster_set_ha = no
handy_ha_needed = no
sed -i "s/^multi_cluster_set_ha.*\+=.*/multi_cluster_set_ha = no/"/etc/onestor/onestor.conf
sed -i "s/^handy_ha_needed.*\+=.*/handy_ha_needed = no/" /etc/onestor/onestor.conf
旧方案脚本:sed -i "s/^multi_cluster_set_ha[[:space:]]\+=.*/multi_cluster_set_ha = no/" /etc/onestor/onestor.conf sed -i "s/^handy_ha_needed[[:space:]]\+=.*/handy_ha_needed = no/" /etc/onestor/onestor.conf
因里面的[[:space:]] 代表匹配任意个空格(但不代表0个!!),遇到没有空格的就失败,所以用上面的替换,“.*”代表“”内任意个字符。
详细说明见文章末尾:
原文链接:https://blog.csdn.net/bandaoyu/article/details/120047612
读取配置项的值
获取“public_addr=172.17.31.109”等号左边的名字public_addr
item=`cat /etc/ceph/ceph.conf |grep -E 'public_addr'`&&echo ${item%%=*}
获取“public_addr=172.17.31.109”等号右边的值 72.17.31.109
line=`cat /etc/ceph/ceph.conf |grep -E 'public_addr'`&&echo ${line#*=}
替换/修改
PermitRootLogin no 替换成PermitRootLogin yes
sudo sed -i "s/PermitRootLogin no/PermitRootLogin yes/g" /etc/ssh/sshd_config
在匹配行行首/行尾添加字符(#)
在每一行的行首/尾添加字符串
sed
-i "
s/.*/行首添加内容&行尾添加内容/"
文件名
在匹配行的行首/尾添加字符串
在包含“STATD#u PORT”的行开头添加“#”注释符号
sed -i "/STATD_PORT/s/^/#/" /tmp/file
在与“callout”匹配的行的末尾添加“your text”
sed -i "/callout/s/$/your text/" /tmp/file
在以“STATD”开头的行的 末尾添加“your text”
sed -i "/^STATD/s/$/your text/" /tmp/file
几点说明:
1."^"代表行首,"$"代表行尾
原文链接:https://blog.csdn.net/bandaoyu/article/details/120047612
去掉匹配行行首/行尾字符(#)
去掉行首#,"^"代表行首
sed -i "/have a nice day/ s/^#//" a.txt
# /have a nice day/ 代表匹配,s/^abcdf// 代表行首的abcdf字符串替换为后面两斜杠里的内容,这里为空,即删除。
去掉行尾#,"$"代表行尾
sed -i "/have a nice day/ s/$#//" a.txt
#匹配行前加
我的记法是a = after ,i = in front
在匹配“2222222222”前面添加“3333333333”,我的记法是a = after ,i = in front
sed -i "/2222222222/i\3333333333" test.txt
#匹配行后加
我的记法是a = after ,i = in front
在匹配“2222222222”后面添加“3333333333”,我的记法是a = after ,i = in front
sed -i "/2222222222/a\3333333333" test.txt
#匹配行前后加
加反斜杠只是为了容易区分,可以不用:
sed -i "/allow 361way.com/iallow www.361way.com" the.conf.file
sed -i "/allow 361way.com/aallow www.361way.com" the.conf.file
这就就可以很方便的看出要在某一行前或某一行后加入什么内容 。
删除匹配行
删除匹配到preSql的行
sed -i "/preSql/d" a.txt
(Linux命令之sed-删除匹配行 :https://www.cnblogs.com/yunjiaofeifei/p/14912272.html)
替换/删除匹配的字符之间的内容
#删除<\/schema>和<heartbeat>字符串之间的内容
sed -i "/<\/schema>/,/<heartbeat>/{//!d}" ./schema.xml
#替换dataNode=".*".*rule="sharding-by-intfile" 为dataNode=" " rule="sharding-by-intfile" .*表示0各或任意个字符
#sed用单引号,所以里面的双引号不用转移符号
sed -i 's/dataNode=".*".*rule="sharding-by-intfile"/dataNode=" " rule="sharding-by-intfile"/g' ./schema.xml
原文链接:https://blog.csdn.net/bandaoyu/article/details/120047612
# Delete text between patterns, excluding the lines containing these patterns(不删除所在行):
sed -i "/PATTERN-1/,/PATTERN-2/{//!d}" input.txt
# Delete text between patterns, including the lines containing these patterns(所在行也删除):
sed -i "/PATTERN-1/,/PATTERN-2/d" input.txt
# To delete all the lines after PATTERN, use this
sed -i "/PATTERN/,$d" input.txt
[sed] Delete the lines lying in between two patterns | *NIX Tricks
sed -i "/#hwBegin/,/#hwEnd/d" /etc/my.cnf #就是将#hwBegin行到#hwEnd行之间内容,包括
#hwBegin行和#hwEnd行都删除,-i表示编辑的是文件
#add '-----here is the content---' to CONF_DIR behind the "[mysql]"
#use command 'sed' to add A behind "pattern" :sed 's/pattern/&A/' filename
用命令'sed'在filename文件内匹配的字符"pattern"后面添加内容A
#!/bin/ash
CONF_DIR="/etc/my.cnf"
sed -i 's/\[mysqld\]/&\n \
wait_timeout=2073600\n \
interactive_timeout=2073600\n \
bulk_insert_buffer_size=16M\n \
max_allowed_packet=16M\n/' /etc/my.cnf
sed命令删除特定行号
删除第N行
1 |
|
删除第N~M行
1 |
|
删除shell变量表示的行号(配合for等语句使用)
1 |
|
删除最后一行
1 |
|
sed命令删除包含特定字符行
删除包含"xxx"的行
1 |
|
1、^代表行首
2、$代表行尾
3、所有行首增加sed -i 's/^/ABC/' a.txt
4、所有行尾添加sed -i 's/$/XYZ/' a.txt
5、删除首行sed -i '1d' d.txt
6、删除末行sed -i '$d' d.txt
7、第5行添加sed -i '5 r 5.txt' a.txt
8、删除空行sed -i '/^$/d' a.txt
9、剔除空格sed -i 's/[ ]*//g' ~/vip1.txt
10、删除回车符sed -i 's/^M//g' a.txt
11、从fromstart这行下面添加内容sed -i '/fromstart/r 4.txt' 5.txt
12、第一列排序存文件awk '{print $1}' vip1.txt |sort -n > vip2.txt
更多实例:
将配置文件中的ms_cluster_type = async+posix改为ms_cluster_type = async+rdma 或者反过来:
方法1:匹配修改
方法2:删除原配置行,添加新配置行
#ps -ef | grep fio | grep -v grep | awk '{print $2}' | xargs kill -s 9
CONF_PATH='/etc/ceph/ceph.conf'
if [ -z $1 ] ; then
echo "******************"
echo "Please input arg:"
echo "******************"
echo "-----------------------------------"
echo "p or P:change ceph conf to posix"
echo "r or R:change ceph conf to rdma"
echo "-----------------------------------"
exitelse
TYPE=$1
fi
case ${TYPE} in
p|P)
sed -i 's/^ms_cluster_type =.*/ms_cluster_type = async+posix/' ${CONF_PATH}
cat ${CONF_PATH} | grep 'ms_cluster_type'if false; then
sed -i '/ms_cluster_type = async+posix/d' ${CONF_PATH} #delete line
sed -i '/ms_cluster_type = async+rdma/d' ${CONF_PATH} #delete linesed -i '/ms_public_type = async+posix/a\ms_cluster_type = async+posix' ${CONF_PATH} #add line after 'ms_public_type = async+posix'
sed -i '/ms_public_type = async+posix/a\#ms_cluster_type = async+rdma' ${CONF_PATH} #add line after 'ms_public_type = async+posix'
fi;;
r|R)
sed -i 's/^ms_cluster_type =.*/ms_cluster_type = async+rdma/' ${CONF_PATH}
cat ${CONF_PATH} | grep 'ms_cluster_type'
if false; then
sed -i '/ms_cluster_type = async+posix/d' ${CONF_PATH} #delete line
sed -i '/ms_cluster_type = async+rdma/d' ${CONF_PATH} #delete linesed -i '/ms_public_type = async+posix/a\#ms_cluster_type = async+posix' ${CONF_PATH} #add line after 'ms_public_type = async+posix'
sed -i '/ms_public_type = async+posix/a\ms_cluster_type = async+rdma' ${CONF_PATH} #add line after 'ms_public_type = async+posix'
fi
;;
*)
esac
匹配任意个空格
[[:space:]] 字符串内的任意个空格
\*[[:space:]]\+hard 行首到hard之间的任意个空格
hard[[:space:]]\+nofile 字符串hard到字符串nofile之间的任意个空格
[[:space:]] ,如果前面没有其他字符则 \*[[:space:]],[[:space:]]\+xxx表示xxx前面的任意个空格,所以xxx在行首,则\*[[:space:]]\+ 匹配行首的任意个空格
原文链接:https://blog.csdn.net/bandaoyu/article/details/120047612
场景:
* soft nofile 630000
* hard nofile 654300
grep 匹配如上文本,由于最后一列数字是可变的,并且每列之间可能存在任意个空格或制表符
grep -E '^\*[[:space:]]+soft[[:space:]]+nofile|^\*[[:space:]]+hard[[:space:]]+nofile' /etc/security/limits.conf
将匹配的结果sed 替换最后一列数字
sed -n 's/\*[[:space:]]\+soft[[:space:]]\+nofile.*/\* soft nofile 654350/p;s/\*[[:space:]]\+hard[[:space:]]\+nofile.*/\* hard nofile 654350/p' /etc/security/limits.conf
区别,在匹配任意个数空格或制表符时,grep 的’+’ 不需要转义,sed 需要转义
补充:sed -r 可以实现正则表达式匹配
原文链接:https://blog.csdn.net/rockstics/article/details/111563857
按行读取配置获取参数执行命令|shell读取配置文件
conf.txt
172.17.31.53:12345 1 1 10000 10 1024
172.17.31.53:12345 1 1 10000 10 2048
172.17.31.53:12345 1 1 10000 10 4096
172.17.31.53:12345 1 1 10000 10 8192
readconf.sh
xargs -I{} echo {} < ./conf.txt
把echo换成自己的脚本/程序
方法二:
a=('1 2 3' '4 5 6' '7 8 9')
for i in "${a[@]}" ; do
b=($i) #此时b就相当于二维数组里面的一维数组了,然后可以再次遍历
for j in "${b[@]}"; do
#do someting
done
done
shell练习--关于二维数组的实现问题 - 回眸一笑百贱生 - 博客园
#!/bin/bash
unset array
for x in `cat xxx.properties`
{
#字符串截取:从左至右第一个'='之前的内容
#echo ${x%%=*}
#字符串截取:从左至右第一个'='之后的内容
#echo ${x#*=}
array[${#array[@]}]="${x%%=*} ${x#*=}"
}
echo ${#array[@]}
echo ${array[@]}
删除空格
快速去除字符串中的空格:https://blog.csdn.net/kwame211/article/details/81231261
- 删除行首空格
sed ‘s/^[ \t]*//g'
- 删除行末空格
sed ‘s/[ \t]*$//g'
- 删除所有的空格
sed s/[[:space:]]//g
1039 ceph -s
1040 systemctl daemon-reload
1041 ceph -s
1042 systemctl start ceph-mds.target
restart ceph
sudo systemctl restart ceph.target
显示ceph
sudo ps -aux|grep -E 'ceph-/*|dse'|grep -v 'grep'|awk '{print $12,$13,$14,$15,$16,$17}'
显示配置
cat /etc/ceph/ceph.conf |grep -E 'ms_public_io_type|ms_public_type|ms_cluster_type'
改配置项:
改public配置为Posix
sed -i 's/^ms_public_type[[:space:]]\+=.*/ms_public_type = async+posix/' /etc/ceph/ceph.conf
改public配置为rdma
sed -i 's/^ms_public_type[[:space:]]\+=.*/ms_public_type = async+rdma/' /etc/ceph/ceph.conf
改cluster配置为Posix
sed -i 's/^ms_cluster_type[[:space:]]\+=.*/ms_cluster_type = async+posix/' /etc/ceph/ceph.conf
改cluster配置为Rdma
sed -i 's/^ms_cluster_type[[:space:]]\+=.*/ms_cluster_type = async+rdma/' /etc/ceph/ceph.conf
改public_io为rdma
sed -i 's/^ms_public_io_type[[:space:]]\+=.*/ms_public_io_type = async+rdma/' /etc/ceph/ceph.conf
改public_io为posix
sed -i 's/^ms_public_io_type[[:space:]]\+=.*/ms_public_io_type = async+posix/' /etc/ceph/ceph.conf
sed -i 's/^ms_public_type[[:space:]]\+=.*/ms_public_type = async+posix/' /etc/ceph/ceph.conf
sed -i 's/^ms_cluster_type[[:space:]]\+=.*/ms_cluster_type = async+posix/' /etc/ceph/ceph.conf
sed -i 's/^ms_public_io_type[[:space:]]\+=.*/ms_public_io_type = async+posix/' /etc/ceph/ceph.conf
集群:
显示配置
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "cat /etc/ceph/ceph.conf |grep -E 'ms_public_io_type|ms_public_type|ms_cluster_type'"
改配置项:
改public配置为Posix
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "sed -i 's/^ms_public_type[[:space:]]\+=.*/ms_public_type = async+posix/' /etc/ceph/ceph.conf"
改public配置为rdma
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "sed -i 's/^ms_public_type[[:space:]]\+=.*/ms_public_type = async+rdma/' /etc/ceph/ceph.conf"
改cluster配置为Posix
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "sed -i 's/^ms_cluster_type[[:space:]]\+=.*/ms_cluster_type = async+posix/' /etc/ceph/ceph.conf"
改cluster配置为Rdma
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "sed -i 's/^ms_cluster_type[[:space:]]\+=.*/ms_cluster_type = async+rdma/' /etc/ceph/ceph.conf"
显示ceph
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "sudo ps -aux|grep -E 'ceph-/*|dse'|grep -v 'grep';echo '=============================================================================='"
restart ceph
line=`cat /etc/ceph/ceph.conf |grep -E 'mon_host'`&&HOSTS=$(echo ${line#*=}|sed s/[[:space:]]//g)&&sudo clush -l root -w ${HOSTS} "sudo systemctl restart ceph.target"
如何在sed中使用变量
第一
在sed条件中是不认识变量取值的
sed '/$x/d' test
所以要想它能够识别变量
sed "/$x/d/" test
方法简单就是把"单引号"变成"双引号"
第二
teststr="IBM"
sed -n '/' "$teststr" '/=' testfile.txt
通常,我们使用sed进行变量替换的时候,替换和被替换变量都是hard-coded的。例如:
sed -n '/comm/p' /tmp/test.log
如果我们用一变量var,它的值根据上下文变化
$ var="comm",定义了变量,那么我们在sed的使用中这样使用变量
$ sed -n ‘/’”$var”‘/p’ /tmp/test.log
注意,是用单引号包含双引号来引用变量。
另:反引号可直接引用变量
例 KEYWORD_SUM=`cat $RUNNING_PATH/keywordfile.txt|wc -l `
awk 获取指定IP对应的网口
[shell] awk 获取指定IP对应的网口;https://blog.csdn.net/llm_hao/article/details/108654916
IP_ADDRESS=192.168.130.12 INTERFACE_NAME=`ifconfig | awk -F ":" '/'$IP_ADDRESS'/{print a}{a=$1}'`
使用shell获取网口名称
for i in `ls /sys/class/net`
do
echo -n `ip a show "$i" | awk '$1 ~ /inet$/' | awk -F "[/ ]+" '{print $NF," -- ",$3," (IPV4)"}'`
ip a show "$i" | awk '$1 ~ /inet6$/' | awk -F "[/ ]+" '{print " ",$3," (IPV6)"}'
done