自动化部署--shell脚本--2
node1和node2都装apache
[root@linux-node1 ~]# yum install httpd -y
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
* base: mirrors.zju.edu.cn
* epel: mirrors.tuna.tsinghua.edu.cn
* extras: mirrors.zju.edu.cn
* updates: mirrors.163.com
Package httpd-2.4.6-45.el7.centos.4.x86_64 already installed and latest version
Nothing to do
[root@linux-node1 ~]# [root@linux-node2 ~]# yum install httpd -y
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
* base: mirrors.zju.edu.cn
* extras: mirrors.zju.edu.cn
* updates: mirrors.163.com
Package httpd-2.4.6-45.el7.centos.4.x86_64 already installed and latest version
Nothing to do
[root@linux-node2 ~]#
把apache的根目录都改成/opt/webroot
[root@linux-node2 ~]# vim /etc/httpd/conf/httpd.conf
[root@linux-node2 ~]#
上面是默认的，改成如下配置
node1也改成如下
启动apache
[root@linux-node2 ~]# systemctl start httpd
[root@linux-node2 ~]# lsof -i:80
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
httpd 21937 root 4u IPv6 132900 0t0 TCP *:http (LISTEN)
httpd 21938 apache 4u IPv6 132900 0t0 TCP *:http (LISTEN)
httpd 21939 apache 4u IPv6 132900 0t0 TCP *:http (LISTEN)
httpd 21940 apache 4u IPv6 132900 0t0 TCP *:http (LISTEN)
httpd 21941 apache 4u IPv6 132900 0t0 TCP *:http (LISTEN)
httpd 21942 apache 4u IPv6 132900 0t0 TCP *:http (LISTEN)
[root@linux-node2 ~]# [root@linux-node1 ~]# systemctl start httpd
[root@linux-node1 ~]# lsof -i:80
COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME
httpd 23765 root 4u IPv6 117035 0t0 TCP *:http (LISTEN)
httpd 23767 apache 4u IPv6 117035 0t0 TCP *:http (LISTEN)
httpd 23768 apache 4u IPv6 117035 0t0 TCP *:http (LISTEN)
httpd 23769 apache 4u IPv6 117035 0t0 TCP *:http (LISTEN)
httpd 23770 apache 4u IPv6 117035 0t0 TCP *:http (LISTEN)
httpd 23771 apache 4u IPv6 117035 0t0 TCP *:http (LISTEN)
[root@linux-node1 ~]#
访问网页
修改首页
[www@linux-node1 ~]$ cd /deploy/code/
[www@linux-node1 code]$ cd web-demo/
[www@linux-node1 web-demo]$ pwd
/deploy/code/web-demo
[www@linux-node1 web-demo]$ ll
total 4
-rw-r--r-- 1 www www 9 Apr 26 22:28 index.html
[www@linux-node1 web-demo]$ cat index.html
nmap.com
[www@linux-node1 web-demo]$ cd /scripts/
[www@linux-node1 scripts]$
再次部署
[www@linux-node1 scripts]$ ./deploy.sh deploy
git pull
code_build
code_scp
web-demo_456_2017-04-27-21-40-36.tar.gz 100% 220 0.2KB/s 00:00
web-demo_456_2017-04-27-21-40-36.tar.gz 100% 220 0.2KB/s 00:00
code_deploy
192.168.58.12.crontab.xml 100% 21 0.0KB/s 00:00
code_test
cluster_node_in
[www@linux-node1 scripts]$
检查部署结果，注意ctrl+f5
查看此时自动化部署脚本的配置
[root@linux-node1 scripts]# cat deploy.sh
#!/bin/bash #Dir List
#mkdir -p /deploy/code/web-demo/
#mkdir -p /deploy/config/web-demo/base
#mkdir -p /deploy/config/web-demo/other
#mkdir -p /deploy/tar
#mkdir -p /deploy/tmp
#mkdir -p /opt/webroot
#mkdir -p /webroot
#chown -R www:www /deploy
#chown -R www:www /opt/webroot
#chown -R www:www /webroot #Node List
NODE_LIST="192.168.58.11 192.168.58.12" # Date/Time Veriables
LOG_DATE='date "+%Y-%m-%d"'
LOG_TIME='date "+%H-%M-%S"' CDATE=$(date "+%Y-%m-%d")
CTIME=$(date "+%H-%M-%S") #Shell Env
SHELL_NAME="deploy.sh"
SHELL_DIR="/home/www"
SHELL_LOG="${SHELL_DIR}/${SHELL_NAME}.log" #Code Env
PRO_NAME="web-demo"
CODE_DIR="/deploy/code/web-demo"
CONFIG_DIR="/deploy/config/web-demo"
TMP_DIR="/deploy/tmp"
TAR_DIR="/deploy/tar"
LOCK_FILE="/tmp/deploy.lock" usage(){
echo $"Usage: $0 [ deploy | rollback ]"
} writelog(){
LOGINFO=$1
echo "${CDATE} ${CTIME}: ${SHELL_NAME} : ${LOGINFO}" >> ${SHELL_LOG}
} shell_lock(){
touch ${LOCK_FILE}
}
shell_unlock(){
rm -f ${LOCK_FILE}
} code_get(){
writelog "code_get";
cd $CODE_DIR && echo "git pull"
cp -r ${CODE_DIR} ${TMP_DIR}/
API_VER="456"
} code_build(){
echo code_build
} code_config(){
writelog "code_config"
/bin/cp -r ${CONFIG_DIR}/base/* ${TMP_DIR}/"${PRO_NAME}"
PKG_NAME="${PRO_NAME}"_"${API_VER}"_"${CDATE}-${CTIME}"
cd ${TMP_DIR} && mv ${PRO_NAME} ${PKG_NAME}
} code_tar(){
writelog "code_tar"
cd ${TMP_DIR} && tar cfz ${PKG_NAME}.tar.gz ${PKG_NAME}
writelog "${PKG_NAME}.tar.gz"
} code_scp(){
echo "code_scp"
for node in $NODE_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
} cluster_node_remove(){
writelog "cluster_node_remove"
} code_deploy(){
echo code_deploy
for node in $NODE_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
} code_test(){
echo code_test
} cluster_node_in(){
echo cluster_node_in
} rollback(){
echo rollback
} main(){
if [ -f ${LOCK_FILE} ];then
echo "Deploy is running" && exit;
fi
DEPLOY_METHOD=$1
case $DEPLOY_METHOD in
deploy)
shell_lock;
code_get;
code_build;
code_config;
code_tar;
code_scp;
cluster_node_remove;
code_deploy;
code_test;
cluster_node_in;
shell_unlock;
;;
rollback)
shell_lock;
rollback;
shell_unlock;
;;
*)
usage;
esac }
main $1
[root@linux-node1 scripts]#
修改测试部分,完善测试函数
过滤到，返回0，过滤不到返回其它的
-s静默模式
[root@linux-node1 scripts]# curl --head http://192.168.58.11/index.html | grep "200 OK"
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 9 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
HTTP/1.1 200 OK
[root@linux-node1 scripts]# curl -s --head http://192.168.58.11/index.html | grep "200 OK"
HTTP/1.1 200 OK
[root@linux-node1 scripts]#
准备是这样的，部署一个，测试一个，通了才继续往下部署。这样才合理
很多公司拿预热节点部署。通过之后再部署生产节点
先修改下上面，把主机分组，模拟分组部署
GROUP1_LIST模拟预热节点
GROUP2_LIST 模拟其余的全部节点
#chown -R www:www /opt/webroot
#chown -R www:www /webroot #Node List
GROUP1_LIST="192.168.58.11"
GROUP2_LIST="192.168.58.12" # Date/Time Veriables
LOG_DATE='date "+%Y-%m-%d"'
LOG_TIME='date "+%H-%M-%S"'
scp这里改下
code_scp(){
writelog "code_scp"
for node in $GROUP1_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
for node in $GROUP2_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
}
把code_deploy也修改下
下面这里是修改之前
cluster_node_remove(){
writelog "cluster_node_remove"
}
code_deploy(){
echo code_deploy
for node in $NODE_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
}
code_test(){
echo code_test
}
修改之后
cluster_node_remove(){
writelog "cluster_node_remove"
}
group1_deploy(){
echo code_deploy
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
}
group2_deploy(){
echo code_deploy
for node in $GROUP2_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
}
cluster_node_in(){
echo cluster_node_in
}
第一组弄的只有1台机器，作为预生产节点
一个节点如果重启需要1分钟的话，5个节点岂不是5分钟。
所以，1个节点先作为预生产节点，部署完毕，测试完毕之后，其它节点可以直接部署了。思想上是这样
添加group1_test测试函数
group1_deploy(){
echo code_deploy
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
}
group1_test(){
curl -s --head http://192.168.58.11/index.html | grep "200 OK"
if [ $? -ne 0 ];then
shell_unlock;
writelog "test error" && exit;
fi
}
group2_deploy(){
echo code_deploy
for node in $GROUP2_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
}
由于group2也要测试，并且每个节点都要测试
因此可以把测试的部分提取出来，封装成函数
shell_lock(){
touch ${LOCK_FILE}
}
url_test(){
URL=$1
curl -s --head $URL | grep "200 OK"
if [ $? -ne 0 ];then
shell_unlock;
writelog "test error" && exit;
fi
}
shell_unlock(){
rm -f ${LOCK_FILE}
}
测试部分可以这么写
cluster_node_remove(){
writelog "cluster_node_remove"
}
group1_deploy(){
echo “code_deploy”
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
}
group1_test(){
curl_test "http://192.168.58.11/index.html"
echo "add to cluster"
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
}
group1_test(){
url_test "http://192.168.58.11/index.html"
echo "add to cluster"
}
group2_deploy(){
echo “code_deploy”
for node in $GROUP2_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
}
group2_test(){
url_test "http://192.168.58.12/index.html"
echo "add to cluster"
}
cluster_node_in(){
echo cluster_node_in
}
可以把下面函数删除。是否可以加入集群，放在测试函数即可
cluster_node_in(){
echo cluster_node_in
}
主函数会这么写
cluster_node_remove(){
writelog "cluster_node_remove"
DEPLOY_METHOD=$1
case $DEPLOY_METHOD in
deploy)
shell_lock;
code_get;
code_build;
code_config;
code_tar;
code_scp;
group1_deploy;
group1_test;
group2_deploy;
group2_test;
shell_unlock;
;;
rollback)
shell_lock;
rollback;
shell_unlock;
;;
*)
usage;
esac
}
main $1
测试脚本
[www@linux-node1 scripts]$ ./deploy.sh deploy
git pull
code_build
web-demo_456_2017-04-27-22-41-54.tar.gz 100% 214 0.2KB/s 00:00
web-demo_456_2017-04-27-22-41-54.tar.gz 100% 214 0.2KB/s 00:00
code_deploy
HTTP/1.1 200 OK
add to cluster
code_deploy
192.168.58.12.crontab.xml 100% 21 0.0KB/s 00:00
HTTP/1.1 200 OK
add to cluster
[www@linux-node1 scripts]$
没问题，可以访问
继续优化脚本
pre是预热节点，生产只有1个
group1_list里面有多个
#Node List
PRE_LIST="192.168.58.11"
GROUP1_LIST="192.168.58.12" # Date/Time Veriables
LOG_DATE='date "+%Y-%m-%d"'
LOG_TIME='date "+%H-%M-%S"' CDATE=$(date "+%Y-%m-%d")
CTIME=$(date "+%H-%M-%S")
下面这里改下
code_scp(){
writelog "code_scp"
for node in $PRE_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
for node in $GROUP1_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
}
上面的remove函数不需要了。删除即可，如下
cluster_node_remove(){
writelog "cluster_node_remove"
}
下面的group_list 名字还需要改，再次优化下，把原先的group1_list改成pre_test，然后下面这里的改成group1
下面的group1_test应该改成for循环，测试里面的节点，可以测一个加一个节点到集群。还可以测试完，一块加测试的时候，
如果你只有2-3个节点，没必要写for循环
优化之后内容
[root@linux-node1 scripts]# sed -n '89,128p' deploy.sh
code_scp(){
writelog "code_scp"
for node in $PRE_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
for node in $GROUP1_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
} pre_deploy(){
writelog "remove from cluster"
ssh $PRE_LIST "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $PRE_LIST "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
} pre_test(){
url_test "http://${PRE_LIST}/index.html"
echo "add to cluster"
} group1_deploy(){
writelog "remove from cluster"
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
} group1_test(){
url_test "http://192.168.58.12/index.html"
echo "add to cluster"
} rollback(){
echo rollback
}
[root@linux-node1 scripts]#
测试的时候，如果只有几个节点，直接几行地址就行了嘛。没必要写太复杂
group1_test(){
url_test "http://192.168.58.12/index.html"
url_test "http://192.168.58.12/index.html"
url_test "http://192.168.58.12/index.html"
echo "add to cluster"
}
主函数部分改成下面
[root@linux-node1 scripts]# sed -n '130,$p' deploy.sh
main(){
if [ -f ${LOCK_FILE} ];then
echo "Deploy is running" && exit;
fi
DEPLOY_METHOD=$1
case $DEPLOY_METHOD in
deploy)
shell_lock;
code_get;
code_build;
code_config;
code_tar;
code_scp;
pre_deploy;
pre_test;
group1_deploy;
group1_test;
shell_unlock;
;;
rollback)
shell_lock;
rollback;
shell_unlock;
;;
*)
usage;
esac }
main $1
[root@linux-node1 scripts]#
改下代码，区分下主页显示和之前的不同
[www@linux-node1 scripts]$ cd /deploy/code/web-demo/
[www@linux-node1 web-demo]$ cat index.html
nmap.com
[www@linux-node1 web-demo]$ echo "www.nmap.com" > index.html
[www@linux-node1 web-demo]$ cat index.html
www.nmap.com
[www@linux-node1 web-demo]$
再次部署
[www@linux-node1 scripts]$ ./deploy.sh deploy
git pull
code_build
web-demo_456_2017-04-27-23-17-27.tar.gz 100% 220 0.2KB/s 00:00
web-demo_456_2017-04-27-23-17-27.tar.gz 100% 220 0.2KB/s 00:00
HTTP/1.1 200 OK
add to cluster
192.168.58.12.crontab.xml 100% 21 0.0KB/s 00:00
HTTP/1.1 200 OK
add to cluster
[www@linux-node1 scripts]$
检查页面，没问题
此时脚本内容如下
[root@linux-node1 scripts]# cat deploy.sh
#!/bin/bash #Dir List
#mkdir -p /deploy/code/web-demo/
#mkdir -p /deploy/config/web-demo/base
#mkdir -p /deploy/config/web-demo/other
#mkdir -p /deploy/tar
#mkdir -p /deploy/tmp
#mkdir -p /opt/webroot
#mkdir -p /webroot
#chown -R www:www /deploy
#chown -R www:www /opt/webroot
#chown -R www:www /webroot #Node List
PRE_LIST="192.168.58.11"
GROUP1_LIST="192.168.58.12" # Date/Time Veriables
LOG_DATE='date "+%Y-%m-%d"'
LOG_TIME='date "+%H-%M-%S"' CDATE=$(date "+%Y-%m-%d")
CTIME=$(date "+%H-%M-%S") #Shell Env
SHELL_NAME="deploy.sh"
SHELL_DIR="/home/www"
SHELL_LOG="${SHELL_DIR}/${SHELL_NAME}.log" #Code Env
PRO_NAME="web-demo"
CODE_DIR="/deploy/code/web-demo"
CONFIG_DIR="/deploy/config/web-demo"
TMP_DIR="/deploy/tmp"
TAR_DIR="/deploy/tar"
LOCK_FILE="/tmp/deploy.lock" usage(){
echo $"Usage: $0 [ deploy | rollback ]"
} writelog(){
LOGINFO=$1
echo "${CDATE} ${CTIME}: ${SHELL_NAME} : ${LOGINFO}" >> ${SHELL_LOG}
} shell_lock(){
touch ${LOCK_FILE}
} url_test(){
URL=$1
curl -s --head $URL | grep "200 OK"
if [ $? -ne 0 ];then
shell_unlock;
writelog "test error" && exit;
fi
} shell_unlock(){
rm -f ${LOCK_FILE}
} code_get(){
writelog "code_get";
cd $CODE_DIR && echo "git pull"
cp -r ${CODE_DIR} ${TMP_DIR}/
API_VER="456"
} code_build(){
echo code_build
} code_config(){
writelog "code_config"
/bin/cp -r ${CONFIG_DIR}/base/* ${TMP_DIR}/"${PRO_NAME}"
PKG_NAME="${PRO_NAME}"_"${API_VER}"_"${CDATE}-${CTIME}"
cd ${TMP_DIR} && mv ${PRO_NAME} ${PKG_NAME}
} code_tar(){
writelog "code_tar"
cd ${TMP_DIR} && tar cfz ${PKG_NAME}.tar.gz ${PKG_NAME}
writelog "${PKG_NAME}.tar.gz"
} code_scp(){
writelog "code_scp"
for node in $PRE_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
for node in $GROUP1_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
} pre_deploy(){
writelog "remove from cluster"
ssh $PRE_LIST "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $PRE_LIST "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
} pre_test(){
url_test "http://${PRE_LIST}/index.html"
echo "add to cluster"
} group1_deploy(){
writelog "remove from cluster"
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
} group1_test(){
url_test "http://192.168.58.12/index.html"
echo "add to cluster"
} rollback(){
echo rollback
} main(){
if [ -f ${LOCK_FILE} ];then
echo "Deploy is running" && exit;
fi
DEPLOY_METHOD=$1
case $DEPLOY_METHOD in
deploy)
shell_lock;
code_get;
code_build;
code_config;
code_tar;
code_scp;
pre_deploy;
pre_test;
group1_deploy;
group1_test;
shell_unlock;
;;
rollback)
shell_lock;
rollback;
shell_unlock;
;;
*)
usage;
esac }
main $1
[root@linux-node1 scripts]#
自动化部署实战-秒级回滚
再加个能紧急回滚的，就不考虑用户体验了。以紧急回滚为主。也不测试了
更紧急流程。直接执行回滚上一个版本
先写个第一种正常流程的，列出回滚版本的函数
因此ROLLBACK_VER=$2 和rollback $ROLLBACK_VER ;
[root@linux-node1 scripts]# sed -n '129,$p' deploy.sh
main(){
if [ -f ${LOCK_FILE} ];then
echo "Deploy is running" && exit;
fi
DEPLOY_METHOD=$1
ROLLBACK_VER=$2
case $DEPLOY_METHOD in
deploy)
shell_lock;
code_get;
code_build;
code_config;
code_tar;
code_scp;
pre_deploy;
pre_test;
group1_deploy;
group1_test;
shell_unlock;
;;
rollback)
shell_lock;
rollback $ROLLBACK_VER ;
shell_unlock;
;;
*)
usage;
esac }
main $1
[root@linux-node1 scripts]#
完善rollback函数
group1_test(){
url_test "http://192.168.58.12/index.html"
echo "add to cluster"
}
rollback_fun(){
for node in $ROLLBACK_LIST;do
ssh $node "rm -f /webroot/web-demo && ln-s /opt/webroo/$1 /webroot/web-demo"
done
}
rollback(){
case $1 in
list)
ls -l /opt/webroot/*.tar.gz
;;
*)
rollback_fun $1
esac
}
main(){
if [ -f ${LOCK_FILE} ];then
最前面加上如下行，然后写上所有节点
#Node List
PRE_LIST="192.168.58.11"
GROUP1_LIST="192.168.58.12"
ROLLBACK_LIST="192.168.58.11 192.168.58.12" # Date/Time Veriables
LOG_DATE='date "+%Y-%m-%d"'
LOG_TIME='date "+%H-%M-%S"'
下面这里改成如下
usage(){
echo $"Usage: $0 { deploy | rollback [ list |version ]}"
}
上面写不合适，应该不传参数列出来可以回退的版本
现在如果rollback一个不存在的版本，它会把软链接删除了，回退也失败，因此可以反过来
再改下
之所以把这个if判断写在for下面。主要是，如果只部署了预生产节点，没部署其它节点
那么回退的时候就可以把预部署节点回退了。否则其它的都回退失败了，并且还要检测其它节点上有没有包。我觉得这里应该ssh -d检测
rollback_fun(){
for node in $ROLLBACK_LIST;do
if [ -d /opt/webroot/$1 ];then
ssh $node "rm -f /webroot/web-demo && ln -s /opt/webroo/$1 /webroot/web-demo"
fi
done
}
继续改造（其实这里通过ssh之后判断其余节点目录是否存在的）
rollback_fun(){
for node in $ROLLBACK_LIST;do
if [ -d /opt/webroot/$1 ];then
ssh $node "[ -d /opt/webroot/$1 ] && rm -f /webroot/web-demo && ln -s /opt/webroo/$1 /webroot/web-demo"
fi
done
}
这里就不对了吧。rollback之后没跟参数居然执行了
[www@linux-node1 scripts]$ ./deploy.sh rollback
[www@linux-node1 scripts]$
改造下
rollback_fun(){
for node in $ROLLBACK_LIST;do
if [ -d /opt/webroot/$1 ];then
ssh $node "if [ -d /opt/webroot/$1 ];then rm -f /webroot/web-demo && ln -s /opt/webroo/$1 /webroot/web-demo;else echo "hehe";fi"
fi
done
}
执行
[www@linux-node1 scripts]$ ./deploy.sh rollback
[www@linux-node1 scripts]$
上面执行为和之前显示一致，再改造下。
rollback_fun(){
for node in $ROLLBACK_LIST;do
ssh $node "if [ -d /opt/webroot/$1 ];then rm -f /webroot/web-demo && ln -s /opt/webroo/$1 /webroot/web-demo;else echo "hehe";fi"
done
}
测试
[www@linux-node1 scripts]$ ./deploy.sh rollback
[www@linux-node1 scripts]$
执行还是没达到效果
先妥协了。以下面为准。虽然如果少参数，会导致执行不成功，先这样
rollback_fun(){
for node in $ROLLBACK_LIST;do ssh $node "rm -f /webroot/web-demo && ln -s /opt/webroo/$1 /webroot/web-demo"
done
}
注意把脚本最后一行改成main $1 $2
[root@linux-node1 scripts]# tail -10 deploy.sh
shell_lock;
rollback ${ROLLBACK_VER};
shell_unlock;
;;
*)
usage;
esac }
main $1 $2
[root@linux-node1 scripts]#
继续测试，修改首页内容
[www@linux-node1 web-demo]$ cd /deploy/code/web-demo
[www@linux-node1 web-demo]$ ls
index.html
[www@linux-node1 web-demo]$ cat index.html
www.nmap.com
[www@linux-node1 web-demo]$ echo 'www.nmap.org' >index.html
[www@linux-node1 web-demo]$ cat index.html
www.nmap.org
[www@linux-node1 web-demo]$
继续执行脚本部署
[www@linux-node1 web-demo]$ cd /scripts/
[www@linux-node1 scripts]$ ./deploy.sh deploy
git pull
code_build
web-demo_456_2017-04-28-21-17-17.tar.gz 100% 222 0.2KB/s 00:00
web-demo_456_2017-04-28-21-17-17.tar.gz 100% 222 0.2KB/s 00:00
HTTP/1.1 200 OK
add to cluster
192.168.58.12.crontab.xml 100% 21 0.0KB/s 00:00
HTTP/1.1 200 OK
add to cluster
[www@linux-node1 scripts]$
检查页面
列出回滚版本
[www@linux-node1 scripts]$ ./deploy.sh rollback list
-rw-rw-r-- 1 www www 204 Apr 23 23:33 /opt/webroot/web-demo_123_2017-04-23-23-33-50.tar.gz
-rw-rw-r-- 1 www www 204 Apr 23 23:43 /opt/webroot/web-demo_123_2017-04-23-23-43-48.tar.gz
-rw-rw-r-- 1 www www 204 Apr 24 00:00 /opt/webroot/web-demo_123_2017-04-24-00-00-14.tar.gz
-rw-rw-r-- 1 www www 204 Apr 24 00:01 /opt/webroot/web-demo_123_2017-04-24-00-01-24.tar.gz
-rw-rw-r-- 1 www www 205 Apr 24 00:02 /opt/webroot/web-demo_123_2017-04-24-00-02-44.tar.gz
-rw-rw-r-- 1 www www 204 Apr 24 00:04 /opt/webroot/web-demo_456_2017-04-24-00-04-05.tar.gz
-rw-rw-r-- 1 www www 220 Apr 27 21:40 /opt/webroot/web-demo_456_2017-04-27-21-40-36.tar.gz
-rw-rw-r-- 1 www www 214 Apr 27 22:41 /opt/webroot/web-demo_456_2017-04-27-22-41-54.tar.gz
-rw-rw-r-- 1 www www 220 Apr 27 23:17 /opt/webroot/web-demo_456_2017-04-27-23-17-27.tar.gz
-rw-rw-r-- 1 www www 220 Apr 28 20:15 /opt/webroot/web-demo_456_2017-04-28-20-15-56.tar.gz
-rw-rw-r-- 1 www www 220 Apr 28 20:16 /opt/webroot/web-demo_456_2017-04-28-20-16-06.tar.gz
-rw-rw-r-- 1 www www 221 Apr 28 20:24 /opt/webroot/web-demo_456_2017-04-28-20-24-23.tar.gz
-rw-rw-r-- 1 www www 220 Apr 28 20:25 /opt/webroot/web-demo_456_2017-04-28-20-25-47.tar.gz
-rw-rw-r-- 1 www www 220 Apr 28 20:28 /opt/webroot/web-demo_456_2017-04-28-20-28-03.tar.gz
-rw-rw-r-- 1 www www 221 Apr 28 20:35 /opt/webroot/web-demo_456_2017-04-28-20-35-37.tar.gz
-rw-rw-r-- 1 www www 222 Apr 28 20:43 /opt/webroot/web-demo_456_2017-04-28-20-43-32.tar.gz
-rw-rw-r-- 1 www www 221 Apr 28 20:45 /opt/webroot/web-demo_456_2017-04-28-20-45-23.tar.gz
-rw-rw-r-- 1 www www 220 Apr 28 20:48 /opt/webroot/web-demo_456_2017-04-28-20-48-01.tar.gz
-rw-rw-r-- 1 www www 222 Apr 28 21:14 /opt/webroot/web-demo_456_2017-04-28-21-14-49.tar.gz
-rw-rw-r-- 1 www www 222 Apr 28 21:17 /opt/webroot/web-demo_456_2017-04-28-21-17-17.tar.gz
[www@linux-node1 scripts]$
回滚上一个版本web-demo_456_2017-04-28-21-14-49
[www@linux-node1 scripts]$ ./deploy.sh rollback web-demo_456_2017-04-28-21-14-49
[www@linux-node1 scripts]$
测试访问
回滚成功，此时脚本内容如下
[root@linux-node1 scripts]# cat deploy.sh
#!/bin/bash #Dir List
#mkdir -p /deploy/code/web-demo/
#mkdir -p /deploy/config/web-demo/base
#mkdir -p /deploy/config/web-demo/other
#mkdir -p /deploy/tar
#mkdir -p /deploy/tmp
#mkdir -p /opt/webroot
#mkdir -p /webroot
#chown -R www:www /deploy
#chown -R www:www /opt/webroot
#chown -R www:www /webroot #Node List
PRE_LIST="192.168.58.11"
GROUP1_LIST="192.168.58.12"
ROLLBACK_LIST="192.168.58.11 192.168.58.12" # Date/Time Veriables
LOG_DATE='date "+%Y-%m-%d"'
LOG_TIME='date "+%H-%M-%S"' CDATE=$(date "+%Y-%m-%d")
CTIME=$(date "+%H-%M-%S") #Shell Env
SHELL_NAME="deploy.sh"
SHELL_DIR="/home/www"
SHELL_LOG="${SHELL_DIR}/${SHELL_NAME}.log" #Code Env
PRO_NAME="web-demo"
CODE_DIR="/deploy/code/web-demo"
CONFIG_DIR="/deploy/config/web-demo"
TMP_DIR="/deploy/tmp"
TAR_DIR="/deploy/tar"
LOCK_FILE="/tmp/deploy.lock" usage(){
echo $"Usage: $0 { deploy | rollback [ list |version ]}"
} writelog(){
LOGINFO=$1
echo "${CDATE} ${CTIME}: ${SHELL_NAME} : ${LOGINFO}" >> ${SHELL_LOG}
} shell_lock(){
touch ${LOCK_FILE}
} url_test(){
URL=$1
curl -s --head $URL | grep "200 OK"
if [ $? -ne 0 ];then
shell_unlock;
writelog "test error" && exit;
fi
} shell_unlock(){
rm -f ${LOCK_FILE}
} code_get(){
writelog "code_get";
cd $CODE_DIR && echo "git pull"
cp -r ${CODE_DIR} ${TMP_DIR}/
API_VER="456"
} code_build(){
echo code_build
} code_config(){
writelog "code_config"
/bin/cp -r ${CONFIG_DIR}/base/* ${TMP_DIR}/"${PRO_NAME}"
PKG_NAME="${PRO_NAME}"_"${API_VER}"_"${CDATE}-${CTIME}"
cd ${TMP_DIR} && mv ${PRO_NAME} ${PKG_NAME}
} code_tar(){
writelog "code_tar"
cd ${TMP_DIR} && tar cfz ${PKG_NAME}.tar.gz ${PKG_NAME}
writelog "${PKG_NAME}.tar.gz"
} code_scp(){
writelog "code_scp"
for node in $PRE_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
for node in $GROUP1_LIST;do
scp ${TMP_DIR}/${PKG_NAME}.tar.gz $node:/opt/webroot/
done
} pre_deploy(){
writelog "remove from cluster"
ssh $PRE_LIST "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $PRE_LIST "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
} pre_test(){
url_test "http://${PRE_LIST}/index.html"
echo "add to cluster"
} group1_deploy(){
writelog "remove from cluster"
for node in $GROUP1_LIST;do
ssh $node "cd /opt/webroot/ && tar xfz ${PKG_NAME}.tar.gz"
ssh $node "rm -rf /webroot/web-demo && ln -s /opt/webroot/${PKG_NAME} /webroot/web-demo"
done
scp ${CONFIG_DIR}/other/192.168.58.12.crontab.xml 192.168.58.12:/webroot/web-demo/crontab.xml
} group1_test(){
url_test "http://192.168.58.12/index.html"
echo "add to cluster"
} rollback_fun(){
for node in $ROLLBACK_LIST;do
ssh $node "rm -f /webroot/web-demo && ln -s /opt/webroot/$1 /webroot/web-demo"
done
} rollback(){
if [ -z $1 ];then
shell_unlock;
echo "Please input rollback version" && exit;
fi
case $1 in
list)
ls -l /opt/webroot/*.tar.gz
;;
*)
rollback_fun $1
esac
} main(){
if [ -f ${LOCK_FILE} ];then
echo "Deploy is running" && exit;
fi
DEPLOY_METHOD=$1
ROLLBACK_VER=$2
case $DEPLOY_METHOD in
deploy)
shell_lock;
code_get;
code_build;
code_config;
code_tar;
code_scp;
pre_deploy;
pre_test;
group1_deploy;
group1_test;
shell_unlock;
;;
rollback)
shell_lock;
rollback ${ROLLBACK_VER};
shell_unlock;
;;
*)
usage;
esac }
main $1 $2
[root@linux-node1 scripts]#
再完善下脚本，rollback的时候，后面参数为空，直接退出。也就是整个脚本的$2 为空，退出脚本
rollback(){
if [ -z $1 ];then
shell_unlock;
echo "Please input rollback version" && exit;
fi
case $1 in
list)
ls -l /opt/webroot/*.tar.gz
;;
*)
rollback_fun $1
esac
}
自动化部署实战-gitlab和部署和回滚
把gitllab装上，参照链接
GitLab是一个利用 Ruby on Rails 开发的开源应用程序，实现一个自托管的Git项目仓库，可通过Web界面进行访问公开的或者私人项目。
GitLab拥有与Github类似的功能，能够浏览源代码，管理缺陷和注释。可以管理团队对仓库的访问，它非常易于浏览提交过的版本并提供一个文件历史库。
它还提供一个代码片段收集功能可以轻松实现代码复用，便于日后有需要的时候进行查找。
1.基础环境准备
[root@linux-node1 ~]# yum install curl policycoreutils openssh-server openssh-clients postfix -y
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
* base: mirrors.163.com
* epel: mirror01.idc.hinet.net
* extras: mirrors.aliyun.com
* updates: mirrors.163.com
Package curl-7.29.0-35.el7.centos.x86_64 already installed and latest version
Package policycoreutils-2.5-11.el7_3.x86_64 already installed and latest version
Package openssh-server-6.6.1p1-35.el7_3.x86_64 already installed and latest version
Package openssh-clients-6.6.1p1-35.el7_3.x86_64 already installed and latest version
Package 2:postfix-2.10.1-6.el7.x86_64 already installed and latest version
Nothing to do
[root@linux-node1 ~]#
启动postfix
[root@linux-node1 ~]# systemctl start postfix
[root@linux-node1 ~]#
2.安装gitlab-ce
注：由于网络问题，国内用户，建议使用163或者阿里云镜像源进行安装：
3.配置并启动gitlab-ce
默认从163的源里下载，清华大学的镜像源makecache时报错
此包286MB，安装时占800MB多。生产中注意下
[root@linux-node1 ~]# yum install gitlab-ce -y
Loaded plugins: fastestmirror
Loading mirror speeds from cached hostfile
* base: mirrors.163.com
* epel: mirrors.ustc.edu.cn
* extras: mirrors.163.com
* updates: mirrors.163.com
Package gitlab-ce-9.1.0-ce.0.el7.x86_64 already installed and latest version
Nothing to do
[root@linux-node1 ~]#
4.配置并启动gitlab-ce
[root@linux-node1 ~]# gitlab-ctl reconfigure
可以使用gitlab-ctl管理gitlab，例如查看gitlab状态：
[root@linux-node1 ~]# gitlab-ctl status
run: gitaly: (pid 46783) 154s; run: log: (pid 46524) 194s
run: gitlab-monitor: (pid 46839) 152s; run: log: (pid 46671) 170s
run: gitlab-workhorse: (pid 46793) 154s; run: log: (pid 46537) 188s
run: logrotate: (pid 46570) 186s; run: log: (pid 46569) 186s
run: nginx: (pid 47163) 2s; run: log: (pid 46544) 187s
run: node-exporter: (pid 46627) 179s; run: log: (pid 46626) 179s
run: postgres-exporter: (pid 46826) 153s; run: log: (pid 46657) 171s
run: postgresql: (pid 46369) 233s; run: log: (pid 46368) 233s
run: prometheus: (pid 46811) 153s; run: log: (pid 46607) 180s
run: redis: (pid 46312) 239s; run: log: (pid 46311) 239s
run: redis-exporter: (pid 46642) 177s; run: log: (pid 46641) 177s
run: sidekiq: (pid 46509) 195s; run: log: (pid 46508) 195s
run: unicorn: (pid 46483) 196s; run: log: (pid 46482) 196s
[root@linux-node1 ~]#
一些常用命令
gitlab-ctl status
gitlab-ctl stop
gitlab-ctl start
gitlab-ctl restart
先停止apache，因为gitlab占用80端口。防止冲突
[root@linux-node1 ~]# systemctl stop httpd
[root@linux-node1 ~]# gitlab-ctl restart
ok: run: gitaly: (pid 47781) 1s
ok: run: gitlab-monitor: (pid 47788) 0s
ok: run: gitlab-workhorse: (pid 47791) 1s
ok: run: logrotate: (pid 47808) 0s
ok: run: nginx: (pid 47814) 1s
ok: run: node-exporter: (pid 47823) 0s
ok: run: postgres-exporter: (pid 47829) 0s
ok: run: postgresql: (pid 47847) 0s
ok: run: prometheus: (pid 47855) 1s
ok: run: redis: (pid 47866) 0s
ok: run: redis-exporter: (pid 47870) 1s
ok: run: sidekiq: (pid 47877) 0s
ok: run: unicorn: (pid 47882) 1s
[root@linux-node1 ~]# netstat -lntp
Active Internet connections (only servers)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 127.0.0.1:9121 0.0.0.0:* LISTEN 47870/redis_exporte
tcp 0 0 127.0.0.1:9090 0.0.0.0:* LISTEN 47855/prometheus
tcp 0 0 127.0.0.1:9187 0.0.0.0:* LISTEN 47829/postgres_expo
tcp 0 0 0.0.0.0:873 0.0.0.0:* LISTEN 1896/rsync
tcp 0 0 127.0.0.1:9100 0.0.0.0:* LISTEN 47823/node_exporter
tcp 0 0 0.0.0.0:111 0.0.0.0:* LISTEN 1/systemd
tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN 47814/nginx: master
tcp 0 0 127.0.0.1:9168 0.0.0.0:* LISTEN 47788/ruby
tcp 0 0 192.168.122.1:53 0.0.0.0:* LISTEN 1346/dnsmasq
tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN 45266/sshd
tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN 1284/master
tcp 0 0 0.0.0.0:8060 0.0.0.0:* LISTEN 47814/nginx: master
tcp 0 0 127.0.0.1:25151 0.0.0.0:* LISTEN 3249/python2
tcp6 0 0 :::873 :::* LISTEN 1896/rsync
tcp6 0 0 :::3306 :::* LISTEN 41106/mysqld
tcp6 0 0 :::111 :::* LISTEN 1/systemd
tcp6 0 0 ::1:9168 :::* LISTEN 47788/ruby
tcp6 0 0 :::22 :::* LISTEN 45266/sshd
tcp6 0 0 ::1:25 :::* LISTEN 1284/master
[root@linux-node1 ~]#
自动化部署--shell脚本--2的更多相关文章
- 自动化部署--shell脚本--1
传统部署方式1.纯手工scp2.纯手工登录git pull .svn update3.纯手工xftp往上拉4.开发给打一个压缩包,rz上去.解压 传统部署缺点:1.全程运维参与,占用大量时间2.上线速 ...
- 自动化部署--shell脚本--3
登录gitlab第一次登录gitlab,需要为root用户修改密码,root用户也是gitlab的超级管理员.http://192.168.58.11设置密码是gitlab 密码不够长,重新设置下,设 ...
- LINUX系统自动化部署shell脚本
#!/bin/shsudo /etc/init.d/tomcatstopwaitsudo rm -rf /opt/tomcat7/work/*waitsudo rm -rf /opt/tomcat7/ ...
- 自动化远程部署shell脚本
历史原因,有一段时间,项目开发采用一种模式:项目开发及代码版本管理在外网,而主要测试在内网.所以为了同步开发进度,每天会将所有服务在外网jenkins上打包好,然后将服务jar包拷进内网,由于内网服务 ...
- 自动化部署shell
yum install curl policycoreutils openssh-server openssh-clients postfix [root@linux-node2 ~]# yum in ...
- Gitee自动化部署python脚本
一.前期准备 1.1 安装环境 1.安装python3 2.打开命令行安装selenium pip install selenium 二.python代码 2.1 源码 #!/usr/bin/pyth ...
- Ambari自动化卸载shell脚本
#!/bin/bash # Program: # uninstall ambari automatic # History: # 2014/01/13 - Ivan - 2862099249@qq.c ...
- shell脚本自动化部署
由于公司技术部团队较小,没有专门的运维团队,所以运维工作技术部承包了. 一.纯人工部署是这样的: 1. 本地打包:一般 maven clean package 2. 借助xftp上传到服务器对应目录 ...
- linux服务器部署jar包以及shell脚本的书写
背景:记录在linux环境下部署jar程序的过程 1 部署过程记录 1.1 程序结构 这里的main函数就在DemRest2.java 文件中. 为了部署方便,要做到以下两点: 1 在导出的jar包中 ...
随机推荐
- Linux文件系统命令 cp
命令名:cp 功能:拷贝文件,把一个文件的内容拷贝到另外一个文件中去. eg: cp source_file dist_file renjg@renjg-HP-Compaq-Pro--MT:~$ cp ...
- IO-MYSQL的理解
数据库IO简介 IO有四种类型:连续读,随机读,随机写和连续写,连续读写的IO size通常比较大(128KB-1MB),主要衡量吞吐量,而随机读写的IO size比较小(小于8KB),主要衡量I ...
- ubantu安装node、npm、cnpm、live-server
更新ubuntu软件源 sudo apt-get update sudo apt-get install -y python-software-properties software-properti ...
- Day18作业及默写
人狗大战 #!/usr/bin/env python # encoding: utf-8 # Author: MeiMeiLong <2559184081@qq.com> # Create ...
- Golang福利爬虫
没事的时候跑两把,穷人专用. package main import ( "bytes" "fmt" "github.com/PuerkitoBio/ ...
- "@P0"附近有语法错误解释及定位修复
解释: "@P0" 指的是sql语句被预编译之后,其所对应的占位符的相对顺序位置,通俗的讲,就是它代表第几个占位符的位置,其中的0,并不是固定的,如果将定位符看做数组的话,这里的数 ...
- SQL注入之Sqli-labs系列第二十六关（过滤空格、注释符、逻辑运算符注入）和第二十六A
开始挑战第二十六关(Trick with comments and space) 0x1看看源代码 (1)过滤了# or and /**/ / \ ,通过判断也过滤了空格 (2)这样一来只能看 ...
- [转]Deep Reinforcement Learning Based Trading Application at JP Morgan Chase
Deep Reinforcement Learning Based Trading Application at JP Morgan Chase https://medium.com/@ranko.m ...
- 文件系统--fs(读)--fs.read
var fs = require('fs');var buf=new Buffer(1024);//fs.open(); //fs.open(path,flags,mode,callback);/* ...
- [转载]PT建站源码（PT服务器原程序）汇总(20100815更新)
Tbsource官方网站(已失效):http://www.tbsource.com/下载地址:http://www.ipv6bbs.com/thread-5152-1-1.html使用站点:CCFbi ...