Python3自动化运维之Fabric模版详解

Posted hwlong

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了Python3自动化运维之Fabric模版详解相关的知识,希望对你有一定的参考价值。

一、概要

  Fabric是基于Python(2.7,3.4+以上版本)实现的SSH命令行工具,简化了SSH的应用程序部署及系统管理任务,它提供了系统基础的操作组件,可以实现本地或远程shell命令,包括命令执行、文件上传、下载及完整执行日志输出等功能。Fabric在paramiko的基础上做了更高一层的封装,操作起来会更加简单。
Fabric官网地址为:http://www.fabfile.org/。

二、Fabric安装

Mac os及CentOS 安装

pip3 install fabric3

Ubuntu安装

[email protected]virtual-machine:~$ python --version
Python 3.6.5
[email protected]-virtual-machine:~$ pip3 install fabric3

详细请参阅官方文档

三、官方提供了一个简单的入门实例

参考官方资料:https://github.com/fabric/fabric/tree/1.13.1

http://www.fabfile.org/changelog-v1.html?highlight=1.x

https://github.com/mathiasertl/fabric/

#!/usr/bin/env python
from fabric.api import run

#定义一个任务函数,通过run方法实现远程执行"uname -s"命令 def host_type(): run(
uname -s)

运行结果

[email protected]virtual-machine:~/devops$ fab -H localhost host_type
[localhost] Executing task host_type
[localhost] run: uname -s
[localhost] Login password for devops:
[localhost] out: Linux
[localhost] out:


Done.
Disconnecting from localhost... done.

其中,fab命令引用文件名fabfile.py,如果使用非默认文件名称,则需要通过"-f"来指定,示例:

fab -H localhost -f host_type.py host_type

 如果目标主机未配置密钥认证信任,将会提示输入目标主机对应账号登录密码。

四、fab常用参数

官方文档:http://docs.fabfile.org/en/1.14/

fab作为Fabric程序的命令行入口,提供了丰富的参数调用,命令格式如下:

fab [options] <command>[:arg1,arg2-val2,host=foo,host=h1;h2...]...

 常用参数介绍。更多参数可使用fab -help查看。

  • -l:显示定义好的任务函数名;
[email protected]virtual-machine:~/devops$ fab -l
Available commands:

    host_type
  • -f:指定fab入口文件,默认入口文件名为fabfile.py;
  • -g:指定网关(中转)设备,比如堡垒机环境,填写堡垒机IP即可;
  • -H:指定目标主机,多态主机用","号分隔;
  • -P:以异步并行方式运行多主机任务,默认为串行运行;
  • -R:指定role(角色),以角色名区分不同业务组设备;
  • -t:设置设备连接超时时间(秒);
  • -T:设置远程主机命令执行超时时间(秒);
  • -w:当命令执行失败,发出警告,而非默认中止任务。

直接使用命令行执行远程操作,示例:

[email protected]virtual-machine:~/devops$ fab -p 1234567 -H localhost -- uname -s
[localhost] Executing task <remainder>
[localhost] run: uname -s
[localhost] out: Linux
[localhost] out:


Done.
Disconnecting from localhost... done.

五、fabfile的编写

 fab命令好似结合我们编写的fabfile.py(其它文件名必须添加-f filename应用)来搭配使用的,部分命令行参数可以通过相应的方法来替代,使之更加灵活,例如"-H 192.168.56.133,192.168.56.134",我们可以通过定义env.hosts来实现,如"env.hosts=[‘192.168.56.133,192.168.56.134‘]"。fabfile的主体由多个自定义的任务函数组成,不同任务实现不同的操作逻辑。

全局属性设定

env对象的作用是定义fabfile的全局设定,支持多个属性,包含目标主机、用户名、密码、等角色,各属性说明如下:

  • evn.host:定义目标主机,可以用IP或主机名表示,以Python的列表形式定义,如evn.hosts[‘192.168.56.133‘,‘192.168.56.134‘]。
  • env.exclude_hosts:排除指定主机,如env.exclude_hosts=[‘192.168.56.133‘]。
  • env.user:定义用户名,如env.user="root"。
  • env.port:定义目标主机端口,默认为22,如env.port="22"。
  • env.password:定义密码,如env.password=‘1234567‘。
  • env.passwords:与password功能一样,区别在于不同主机不同密码的应用场景,需要注意的是,配置passwords是需配置用户、主机、端口等信息,如:
env.passwords = {
    [email protected]:22:1234567,
    [email protected]:22:1234567,
    [email protected]:22:1234567,
    [email protected]:22:1234567
}
  • env.gateway:定义网关(中转、堡垒机)IP,如env.gateway = ‘192.168.56.1‘。
  • env.deploy_release_dir:自定义全局变量,格式:env.+"变量名称",如env.deploy_release_dir、env.age、env.sex等。
  • env.roledefs:定义角色分组,比如web组与db组主机区分开来,定义如下:
env.roledefs = {
    webservers:[192.168.56.131,192.168.56.132,192.168.56.133],
    dbserver:[192.168.56.134,192.168.56.135]
}

引用时使用Python修饰符的形式进行,角色修饰符下面的任务函数为其作用域,下面来看一个示例:

@roles(webservers)
def webtask():
    run(/etc/init.d/nginx start)


@roles(dbservers)
def dbtask():
    run(/etc/init.d/mysql start)


@roles(webservers,dbservers)
def pubclitasj():
    run(uptime)

def deploy():
    execute(webtask)
    execute(dbtask)
    execute(pubclitask)

在命令执行fab deploy就可以实现不同角色执行不同的任务函数了。

常用API

  Fabric提供了一组简单但功能强大的fabric.api命令集,简单地调用这些API就能完成大部分应用场景需求。Fabric常用方法及说明如下:

  • local:执行本地命令,如:local(‘uname -s‘);
  • lcd:切换本地目录,如:lcd(‘/home‘);
  • cd:切换远程目录,如:cd(‘/data/logs‘);
  • run:执行远程命令,如:run(‘free -m‘);
  • sudo:sudo方式执行远程命令,如:sudo(‘/etc/init.d/httpd start‘);
  • put:上传本地文件到远程主机,如:put(‘/home/user.info‘,‘/data/user.info‘);
  • prompt:获得用户输入信息,如:prompt(‘please input user password:‘);
  • confirm:获得提示信息确认,如:confirm("Tests failed. Continue[Y/N]?");
  • reboot:重启远程主机,如:reboot();
  • @task:函数修饰符,标识的函数为fab可调用的,非标记对fab不可见,纯业务逻辑;
  • runs_once:函数修复符,标识的函数只会执行一次,不受多台主机影响。

六、示例一:查看本地与远程主机信息

查看本地信息

本示例调用local()方法执行本地(主控端)命令,添加"@runs_once"修饰符保证该任务函数只执行一次。调用run()方法执行远程命令。

[email protected]virtual-machine:~/devops$ pwd
/home/devops/devops
[email protected]-virtual-machine:~/devops$ cat simple1.py
#!/usr/bin/env python

from fabric.api import *


env.user = ‘devops
env.hosts = [‘localhost]
env.password = 1234567

@runs_once              #查看本地系统信息,当有多台主机时只运行一次
def local_task():       #本地任务函数
    local("uname -a")

通过fab命令调用local_task任务函数运行结果如下:

[email protected]virtual-machine:~/devops$ fab -f simple1.py local_task
[localhost] Executing task local_task
[localhost] local: uname -a
Linux devops-virtual-machine 4.15.0-20-generic #21-Ubuntu SMP Tue Apr 24 06:16:15 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux

Done.

查看远程主机信息

[email protected]virtual-machine:~/devops$ cat simple2.py
#!/usr/bin/env python

from fabric.api import *


env.user = root
env.hosts = [192.168.56.11]
env.password = 1234567



def remote_task():
    with cd(/root):        #"with"的作用是让后面的表达式的语句继承当前状态,实现"cd /root/ && ls -l‘的效果
        run(ls -l)

 调用remote_task任务函数运行结果如下:

[email protected]virtual-machine:~/devops$ fab -f simple2.py remote_task
[192.168.56.11] Executing task remote_task
[192.168.56.11] run: ls -l
[192.168.56.11] out: total 4
[192.168.56.11] out: -rw-------. 1 root root 1273 May 29 11:47 anaconda-ks.cfg
[192.168.56.11] out:


Done.
Disconnecting from 192.168.56.11... done.

七、示例二:动态获取远程目录列表

本示例使用"@task‘修复符标志入口函数go()对外部可见,配合"@runs_once"修饰符接受用户输入,最后调用worktask()任务函数实现远程命令执行。

[email protected]virtual-machine:~/devops$ cat simple3.py
#!/usr/bin/env python

from fabric.api import *

env.user = root
env.hosts = [192.168.56.11,192.168.56.12]
env.password = 1234567


@runs_once           #主机遍历过程中,只有第一台触发此函数
def input_raw():
    return prompt("Please input directory name:",default="/home")


def worktask(dirname):
    run("ls -l "+dirname)


@task           #限定只有go函数对fab命令可见
def go():
    getdirname = input_raw()
    worktask(getdirname)

该示例实现了一个动态输入远程目录名称,再获取目录列表的功能,由于我们只要求输入一次,在显示所有主机上该目录的列表信息,调用一个子函数input_raw()同时配置@runs_once修复符来达到此目的。

执行结果如下:

[email protected]virtual-machine:~/devops$ fab -f simple3.py go
[192.168.56.11] Executing task go
Please input directory name: [/home] /root
[192.168.56.11] run: ls -l /root
[192.168.56.11] out: total 4
[192.168.56.11] out: -rw-------. 1 root root 1273 May 29 11:47 anaconda-ks.cfg
[192.168.56.11] out:

[192.168.56.12] Executing task go
[192.168.56.12] run: ls -l /root
[192.168.56.12] out: total 4
[192.168.56.12] out: -rw-------. 1 root root 1273 May 29 11:59 anaconda-ks.cfg
[192.168.56.12] out:


Done.
Disconnecting from 192.168.56.11... done.
Disconnecting from 192.168.56.12... done.

八、示例三:网关模式文件上传与执行

本示例通过Fabric的env对象定义网关模式,即俗称的中转、堡垒机环境。定义格式为"env.gateway=‘192.168.56.11‘",其中IP“192.168.56.11”为堡垒机IP,再结合任务韩素实现目标主机文件上传与执行的操作。

[email protected]virtual-machine:/home/install$ cat /home/devops/devops/simple4.py
#!/usr/bin/env python

from fabric.api import *
from fabric.context_managers import *
from fabric.contrib.console import confirm


env.user = root
env.gateway = 192.168.56.11                           #定义堡垒机IP,作为文件上传、执行的中转设备
env.hosts = [192.168.56.12,192.168.56.13]

env.passwords = {
    [email protected]:22:1234567,                  #堡垒机账号信息
    [email protected]:22:1234567,
    [email protected]:22:1234567
}

l_pack_path = "/home/install/nginx-1.6.3.tar.gz"        #本地安装包路径
r_pack_path = "/tmp/install"                            #远程安装包路径


@task
def put_task():
    run("mkdir -p /tmp/install")
    with settings(warn_only=True):
        result = put(l_pack_path,r_pack_path)          #上传安装包
    if result.failed and not confirm("put file failed, Continue[Y/N]?"):
        abort("Aborint file put task!")


@task
def run_task():                    #执行远程命令,安装nginx
    with cd(r_pack_path):
        run("tar -xvf nginx-1.6.3.tar.gz")
        with cd("nginx-1.6.3/"):                     #使用with继续继承/tmp/install目录位置状态
            run("./nginx_install.sh")

@task
def go():       #上传、安装
    put_task()
    run_task()

执行命令fab -f simple4.py go。运行结果如下:

技术分享图片
[email protected]virtual-machine:~/devops$ fab -f simple4.py go
[192.168.56.12] Executing task go
[192.168.56.12] run: mkdir -p /tmp/install
[192.168.56.12] put: /home/install/nginx-1.6.3.tar.gz -> /tmp/install/nginx-1.6.3.tar.gz
[192.168.56.12] run: tar -xvf nginx-1.6.3.tar.gz
.....
.....
.....
[192.168.56.12] out: cp conf/nginx.conf /usr/local/nginx/conf/nginx.conf.default
[192.168.56.12] out: test -d /usr/local/nginx/logs         || mkdir -p /usr/local/nginx/logs
[192.168.56.12] out: test -d /usr/local/nginx/logs ||         mkdir -p /usr/local/nginx/logs
[192.168.56.12] out: test -d /usr/local/nginx/html         || cp -R html /usr/local/nginx
[192.168.56.12] out: test -d /usr/local/nginx/logs ||         mkdir -p /usr/local/nginx/logs
[192.168.56.12] out: make[1]: Leaving directory `/tmp/install/nginx-1.6.3[192.168.56.12] out: nginx: the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
[192.168.56.12] out: nginx: configuration file /usr/local/nginx/conf/nginx.conf test is successful
[192.168.56.12] out:

[192.168.56.13] Executing task go
[192.168.56.13] run: mkdir -p /tmp/install
[192.168.56.13] put: /home/install/nginx-1.6.3.tar.gz -> /tmp/install/nginx-1.6.3.tar.gz
[192.168.56.13] run: tar -xvf nginx-1.6.3.tar.gz
....
....
....
[192.168.56.13] out: cp conf/nginx.conf /usr/local/nginx/conf/nginx.conf.default
[192.168.56.13] out: test -d /usr/local/nginx/logs         || mkdir -p /usr/local/nginx/logs
[192.168.56.13] out: test -d /usr/local/nginx/logs ||         mkdir -p /usr/local/nginx/logs
[192.168.56.13] out: test -d /usr/local/nginx/html         || cp -R html /usr/local/nginx
[192.168.56.13] out: test -d /usr/local/nginx/logs ||         mkdir -p /usr/local/nginx/logs
[192.168.56.13] out: make[1]: Leaving directory `/tmp/install/nginx-1.6.3[192.168.56.13] out: nginx: the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
[192.168.56.13] out: nginx: configuration file /usr/local/nginx/conf/nginx.conf test is successful
[192.168.56.13] out:


Done.
Disconnecting from 192.168.56.11... done.
Disconnecting from 192.168.56.12... done.
Disconnecting from 192.168.56.13... done.
运行结果,默认为串行运行

执行命令fab -Pf simple4.py go。运行结果如下:

技术分享图片
[email protected]virtual-machine:~/devops$ fab -Pf simple4.py go
[192.168.56.12] Executing task go
[192.168.56.13] Executing task go
[192.168.56.12] run: mkdir -p /tmp/install
[192.168.56.13] run: mkdir -p /tmp/install
[192.168.56.12] put: /home/install/nginx-1.6.3.tar.gz -> /tmp/install/nginx-1.6.3.tar.gz
[192.168.56.13] put: /home/install/nginx-1.6.3.tar.gz -> /tmp/install/nginx-1.6.3.tar.gz
[192.168.56.12] run: tar -xvf nginx-1.6.3.tar.gz
....
....
....
[192.168.56.12] out: nginx-1.6.3/html/index.html
[192.168.56.12] out: nginx-1.6.3/README
[192.168.56.12] out: nginx-1.6.3/nginx_install.sh
[192.168.56.12] out: nginx-1.6.3/configure
[192.168.56.12] out:

[192.168.56.12] run: ./nginx_install.sh
[192.168.56.13] run: tar -xvf nginx-1.6.3.tar.gz
[192.168.56.13] out: nginx-1.6.3/
[192.168.56.13] out: nginx-1.6.3/src/
....
....
....
[192.168.56.12] out: make[1]: Leaving directory `/tmp/install/nginx-1.6.3[192.168.56.12] out: nginx: the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
[192.168.56.12] out: nginx: configuration file /usr/local/nginx/conf/nginx.conf test is successful
[192.168.56.12] out:
....
....
...
[192.168.56.13] out: make[1]: Leaving directory `/tmp/install/nginx-1.6.3[192.168.56.13] out: nginx: the configuration file /usr/local/nginx/conf/nginx.conf syntax is ok
[192.168.56.13] out: nginx: configuration file /usr/local/nginx/conf/nginx.conf test is successful
[192.168.56.13] out:
加P参数该为异步并行执行,运行结果

九、示例四:文件打包、上传与校验

我们时常做一些文件包分发的工作,实施步骤一般是先压缩打包,在批量上传至目标服务器,最后做一致性校验。本示例通过put()方法实现文件的上传,通过对比本地与远程主机文件的md5,最终实现文件一致性校验。

[email protected]virtual-machine:~/devops$ cat simple5.py
#!/usr/bin/env python

from fabric.api import *
from fabric.context_managers import *
from fabric.contrib.console import confirm

env.user = root
env.hosts = [192.168.56.12,192.168.56.13]
env.passwords = {
    [email protected]:22:1234567,
    [email protected]:22:1234567,
}


@runs_once
def tar_task():             #本地打包任务函数,只执行一次
    with lcd(/home/devops/devops):
        local("tar -zcf devops.tar.gz  *")


@task
def put_task():                 #上传文件任务函数
    run("mkdir -p /root/devops")
    with cd("/root/devops"):
        with settings(warn_only=True):                  #put(上传)出现异常时继续执行,非终止
            result = put("/home/devops/devops/devops.tar.gz","/root/devops/devops.tar.gz")
        if result.failed and not confirm("put file failed.Continue[Y/N]?"):
            abort("Aborting file put task!")                        #出现异常时,确认用户是否继续,(Y继续)



@task
def check_task():               #校验文件任务函数
    with settings(warn_only=True):
        #本地local命令需要配置capture=True才能捕获返回值
        lmd5 = local("md5sum /home/devops/devops/devops.tar.gz",capture=True).split( )[0]
        rmd5 = run("md5sum /root/devops/devops.tar.gz").split( )[0]
    if lmd5 == rmd5:                #对比本地及远程文件md5信息
        prompt("OK")
    else:
        prompt("ERROR")


@task
def go():
    tar_task()
    put_task()
    check_task()

执行命令fab -f simple5.py go。运行结果如下:(提示此程序不支持-P参数并行执行、如需并行执行,程序需要做调整)

[email protected]virtual-machine:~/devops$ fab -f simple5.py go
[192.168.56.12] Executing task go
[localhost] local: tar -zcf devops.tar.gz  *
[192.168.56.12] run: mkdir -p /root/devops
[192.168.56.12] put: /home/devops/devops/devops.tar.gz -> /root/devops/devops.tar.gz
[localhost] local: md5sum /home/devops/devops/devops.tar.gz
[192.168.56.12] run: md5sum /root/devops/devops.tar.gz
[192.168.56.12] out: a1cf2be82647cbed0d41514bd80373de  /root/devops/devops.tar.gz
[192.168.56.12] out:

OK
[192.168.56.13] Executing task go
[192.168.56.13] run: mkdir -p /root/devops
[192.168.56.13] put: /home/devops/devops/devops.tar.gz -> /root/devops/devops.tar.gz
[localhost] local: md5sum /home/devops/devops/devops.tar.gz
[192.168.56.13] run: md5sum /root/devops/devops.tar.gz
[192.168.56.13] out: a1cf2be82647cbed0d41514bd80373de  /root/devops/devops.tar.gz
[192.168.56.13] out:

OK

Done.
Disconnecting from 192.168.56.12... done.
Disconnecting from 192.168.56.13... done.

十、示例五:部署LNMP业务服务环境

本示例通过env.roledefs定义不同主机角色,在使用"@roles(‘webservers‘)"修复符绑定到对应的任务函数,实现不同角色主机的部署差异。

[email protected]virtual-machine:~/devops$ cat simple6.py
#!/usr/bin/env python

from fabric.colors import *
from fabric.api import *

env.user = root
env.roledefs = {
    webservers:[192.168.56.11,192.168.56.12],
    dbservers:[192.168.56.13]
}

env.passwords = {
    [email protected]:22:1234567,
    [email protected]:22:1234567,
    [email protected]:22:1234567,
}

@roles(webservers)                      #使用webtask任务函数引用webservers角色修复符
def webtask():
    print(yellow(Install nginx php php-fpm...))
    with settings(warn_only=True):
        run("yum -y install nginx")
        run("yum -y install php-fpm php-mysql php-mbstring php-xml php-mcrypt php-gd")
        run("chkconfig --levels 235 php-fpm on")
        run("chkconfig --levels 235 nginx on")


@roles(dbservers)                       #dbtask任务函数引用dbservers角色修复符
def dbtask():
    print(yellow("Install Mysql..."))
    with settings(warn_only=True):
        run("yum -y install mysql mysql-server")
        run("chkconfig --levels 235 mysqld on")


@roles(webservers,dbservers)           #publictask任务函数同时引用两个角色修复符
def publictask():                          #部署公共类环境,如epel、ntp等
    print(yellow("Install epel ntp...."))
    with settings(warn_only=True):
        run("wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo")
        run("yum -y install ntp")

def deploy():
    execute(publictask)
    execute(webtask)
    execute(dbtask)

执行命令fab -Pf simple6.py deploy

[email protected]virtual-machine:~/devops$ fab -Pf simple6.py deploy
[192.168.56.11] Executing task publictask
[192.168.56.12] Executing task publictask
[192.168.56.13] Executing task publictask
Install epel ntp....
[192.168.56.13] run: wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo
Install epel ntp....
[192.168.56.12] run: wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo
Install epel ntp....
[192.168.56.11] run: wget -O /etc/yum.repos.d/epel.repo http://mirrors.aliyun.com/repo/epel-7.repo
[192.168.56.12] out: --2018-06-23 20:32:30--  http://mirrors.aliyun.com/repo/epel-7.repo
[192.168.56.11] out: --2018-06-23 20:32:30--  http://mirrors.aliyun.com/repo/epel-7.repo
[192.168.56.13] out: --2018-06-23 20:32:30--  http://mirrors.aliyun.com/repo/epel-7.repo
....
[192.168.56.13] run: yum -y install ntp
[192.168.56.12] run: yum -y install ntp
[192.168.56.11] run: yum -y install ntp
....
....
....
[192.168.56.11] Executing task webtask
[192.168.56.12] Executing task webtask
Install nginx php php-fpm...
[192.168.56.11] run: yum -y install nginx
Install nginx php php-fpm...
[192.168.56.12] run: yum -y install nginx
....
....
....
[192.168.56.13] Executing task dbtask
Install Mysql...
[192.168.56.13] run: rpm -ivh http://dev.mysql.com/get/mysql-community-release-el6-5.noarch.rpm
.....
.....
.....
[192.168.56.13] run: chkconfig --levels 235 mysqld on

Done.

十一、生产环境代码包发布管理

程序生产环境的发布是业务上线的最后一个环境,要求具备源码打包、发布、切换、回滚、版本管理等功能。本示例实现了这一套流程功能,其中版本切换与回滚使用了Linux下的软链接实现。

 

#!/usr/local/env python

from fabric.api import *
from fabric.colors import *
from fabric.context_managers import *
from fabric.contrib.console import confirm
import time

env.user = root
env.host = [192.168.56.12,192.168.56.13]
env.passwords = {
    [email protected]:22:1234567,
    [email protected]:22:1234567,
}

env.project_dev_source = /data/dev/Lwebadmin/              #开发服务器项目主目录
env.project_tar_source = /data/dev/releases/               #开发服务器项目压缩包存储目录
env.project_pack_name = release                            #项目压缩包前缀,文件名为release.tar.gz

env.deploy_project_root = /data/www/Lwebadmin/            #项目生产环境主目录
env.deploy_release_dir = releases                         #项目发布目录,位于主目录下面
env.deploy_current_dir = current                          #对外服务的当前版本软链接
env.deploy_version = time.strftime("%Y%m%d")+"v2"           #版本号

@runs_once
def input_versionid():                                      #获得用户输入的版本号,以便做版本回滚操作
    return prompt("Please input project rollback version ID:",default="")


@task
@runs_once
def tar_source():                                           #打包本地项目主目录,并将压缩包存储到本地压缩包目录
    prompt(yellow("Creating source package...."))
    with lcd(env.project_dev_source):
        local("tar -zcf %s.tar.gz ." %(env.project_tar_source + env.project_pack_name))
    prompt(green("Creating source package success!"))


@task
def put_package():                                          #上传任务函数
    prompt(yellow("Start put package...."))
    with settings(warn_only=True):
        with cd(env.deploy_project_root + env.deploy_release_dir):
            run("mkdir %s" %(env.deploy_version))           #创建版本目录
    env.deploy_full_path = env.deploy_project_root + env.deploy_release_dir + "/" + env.deploy_version
    with settings(warn_only=True):                          #上传项目压缩包至此目录
        result = put(env.project_tar_source + env.project_pack_name + ".tar.gz",env.deploy_full_path)
    if result.failed and not ("put file failed,Continue[Y/N]?"):
        abort("Aborting file put task!")

    with cd(env.deploy_full_path):                          #成功解压后删除压缩包
        run("tar -zxvf %s.tar.gz" %(env.project_pack_name))
        run("rm -rf %s.tar.gz" %(env.project_pack_name))

    print(green("Put & untar package success!"))


@task
def make_symlink():                                         #为当前版本目录做软链接
    print(yellow("update current symlink"))
    env.deploy_full_path = env.deploy_project_root + env.deploy_release_dir + "/" + env.deploy_version
    with settings(warn_only=True):                           #删除软链接,重新创建并指定软链接源目录,新版本生效
        run("rm -rf %s" %(env.deploy_project_root + env.deploy_current_dir))
        run("ln -s %s %s" %(env.deploy_full_path,env.deploy_project_root + env.deploy_current_dir))
    print(green("make symlink success!"))


@task
def rollback():                                             #版本回滚任务函数
    print(yellow("rollback project version"))
    versionid = input_versionid()                           #获取用户输入的回滚版本号
    if versionid == ‘‘:
        abort("Project version ID error,abort!")

    env.deploy_full_path = env.deploy_project_root + env.deploy_release_dir + "/" + versionid
    run("rm -r %s" %(env.deploy_project_root + env.deploy_current_dir))
    run("ln -s %s %s" %(env.deploy_full_path,env.deploy_project_root + env.deploy_current_dir))     #删除软链接,重新创建并指定软链接源目录,新版本生效
    print(green("rollback sucess!"))


@task
def go():               #自动化程序版本发布入口函数
    tar_source()
    put_package()
    make_symlink()

在生产环境中将站点的根目录指向"/data/www/Lwebadmin/current",由于使用Linux软链接做切换,管理员的版本发布、回滚操作用户无感知。

 

以上是关于Python3自动化运维之Fabric模版详解的主要内容,如果未能解决你的问题,请参考以下文章

自动化运维之Ansible剑客参上

自动化运维之Ansible模块篇

自动化运维之Ansible安装及命令使用

自动化运维之Ansible安装及基本模块应用

自动化运维之paramiko详解

2自动化运维之SaltStack远程执行详解