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

https://www.unixhot.com/article/48

 

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 ~]#