六ansible-playbook任务控制

Posted 2021-10-17 哭泣的馒头

一、playbook任务控制

1、playbook错误处理

(1)、error​​​

​ignore_errors:对错误的处理方式

--True表示忽略错误继续执行

--False 表示遇到错误就停止执行 默认False

​

cat error_test.yaml

---
- hosts: webserver
  remote_user: root
  tasks:
    - shell: /usr/bin/false
      #ignore_errors: True
    - ping:

​ansible-playbook error_test.yaml -i hosts

报错以后不再执行下面的ping

cat error_test.yaml

---
- hosts: webserver
  remote_user: root
  tasks:
    - shell: /usr/bin/false
      ignore_errors: True
    - ping:

​​ansible-playbook error_test.yaml -i hosts

报错了依旧执行下面的ping

(2)、force_handler

通常情况下，当tasks失败，play将会终止，任何在前面已经被tasks notify的handlers都不会被执行，如果在play中设置了force_handlers:yes ,被通知的handlers就会强制执行

- hosts: webserver
  remote_user: root
  force_handlers: yes ##强制调用handlers
  tasks:

2、tags

​打标签，单独运行某个标签或者忽略某个标签

将webserver机器的httpd.cnf拿过来
scp root@192.168.10.128:/etc/httpd/conf/httpd.conf .

cat httpd.conf |grep 8089
Listen 8089 ##配置文件就改了一下端口

 cat apache.yaml
---
- hosts: webserver
  remote_user: root
  tasks:
    - name: install the latest version of apache
      yum:
        name: httpd
        state: latest
    - copy:
        src: httpd.conf
        dest: /etc/httpd/conf/httpd.conf
        owner: apache
        group: apache
        mode: \'0644\'
      tags: testtag ##此处打了个标签
    - copy:
        src: index.html
        dest: /var/www/html/index.html
        owner: apache
        group: apache
        mode: \'0644\'
    - service:
        name: httpd
        state: started
        enabled: yes

​ansible-playbook apache.yaml -i hosts -t testtag

ansible-playbook apache.yaml -i hosts -t testtag

或者ansible-playbook apache.yaml -i hosts --tags=testtag

如果想忽略某个tags

ansible-playbook apache.yaml -i hosts --skip-tags=testtag

3、handlers

​当关注的资源发生变化时采取的操作

notify这个action可用于在每个play的最后被触发，这样可以避免有多次改变发生时，每次都执行指定的操作，取而代之仅在所有的变化发生完成以后一次性的执行指定操作

​

无论多少个task通知了相同的handler，handler仅会在所有tasks结束后运行一次
只有tasks发生改变了才会通知handler，没有改变则不会触发handler
不能使用handler替代tasks，因为handler是一个特殊的tasks

​在notify中列出的操作成为handler，即notify调用handler中定义的操作

在2中虽然使用标签单独更换了配置文件，但是没有重启，服务端口依旧使用的8080

cat httpd.conf |grep 8090 
Listen 8090 ##刚才试验tag，如果不改端口，两边都是8089，copy过程相当于已经做过不会重复做，不会触发触发器

 cat apache.yaml
---
- hosts: webserver
  remote_user: root
  tasks:
    - name: install the latest version of apache
      yum:
        name: httpd
        state: latest
    - copy:
        src: httpd.conf
        dest: /etc/httpd/conf/httpd.conf
        owner: apache
        group: apache
        mode: \'0644\'
      tags: testtag
      notify: apache restart #这个名字必须和触发器定义的名字一致才能找到触发器
    - copy:
        src: index.html
        dest: /var/www/html/index.html
        owner: apache
        group: apache
        mode: \'0644\'
    - service:
        name: httpd
        state: started
        enabled: yes
  handlers:   ##触发器模块
    - name: apache restart
      service:
        name: httpd
        state: restarted

​ansible-playbook apache.yaml -i hosts --tags=testtag

此时可以直接观察数据

4、when

​有些时候需要在满足特定条件后再触发某一项操作，或在特定的条件下终止某个行为，这个时候需要进行条件判断，when正是解决这个问题的最佳选择，远程中的系统变量facts作为when的条件，可以通过setup模块查看

例1

cat httpd.conf |grep 8091
Listen 8091 ###再换个端口

cat apache.yaml
---
- hosts: webserver
  remote_user: root
  tasks:
    - name: install the latest version of apache
      yum:
        name: httpd
        state: latest
    - copy:
        src: httpd.conf
        dest: /etc/httpd/conf/httpd.conf
        owner: apache
        group: apache
        mode: \'0644\'
      when: ansible_os_family == "Debian"  ##此处增加判断，如果操作系统是Debian，就执行这个copy
      tags: testtag
      notify: apache restart
    - copy:
        src: index.html
        dest: /var/www/html/index.html
        owner: apache
        group: apache
        mode: \'0644\'
    - service:
        name: httpd
        state: started
        enabled: yes
  handlers:
    - name: apache restart
      service:
        name: httpd
        state: restarted

​ansible-playbook apache.yaml -i hosts --tags=testtag

​直接忽略了

此时更改when: ansible_os_family == "Debian"为when: ansible_os_family == "RedHat"

​ansible-playbook apache.yaml -i hosts --tags=testtag

例2 根据条件重启服务

[root@db2 memcache]# cat task3.yaml
- hosts: webserver
  tasks:
    - name: Check Httpd Server
      command: systemctl is-active httpd
      ignore_errors: yes  ##如果服务没启动，会报错，这里忽略报错
      register: check_httpd

    - name: Httpd Restarted
      service: name=httpd state=restarted
      when: check_httpd.rc == 0

ansible-playbook task3.yaml -i hosts

5、register

​有时候我们还需要更复杂的例子，比如判断前一个命令的执行结果去处理后面的操作，这时候就需要register模块来保存前一个命令的返回状态，在后面进行调用

cat when_test.yaml

---
- hosts: webserver
  remote_user: root
  tasks:
    - shell: uptime |awk \'{printf("%.2f",$(NF-2))}\'
      register: result
    - debug:
        var: result   #先打印出来result信息，方便查看下面会不会自动关闭apache服务
    - service:
        name: httpd
        state: stopped
      when: result.stdout|float > 0.05 #当负载大于0.05，就关闭apache服务

6、with_items

​是playbook的标准循环，可用于迭代一个列表或字典，通过{{ item }}获取每次迭代的值

例1 批量创建用户

cat user.yaml
---
- hosts: webserver
  remote_user: root
  tasks:
    - name: create user "{{ item.username }}"
      user:
        name: "{{ item.username }}"
        group: "{{ item.group }}"
        password: "{{item.password|password_hash( \'sha512\' )}}"
      with_items:
        -
          username: yantai
          group: shandong
          password: "123456" ###如果为数字，必须加双引号标明为字符串
        -
          username: weihai
          group: shandong
          password: "123456"
        -
          username: weifang
          group: shandong
          password: abc

​ansible-playbook user.yaml -i hosts

例2 重启多个服务

webserver机器已经安装apache和memcached

[root@db2 memcache]# cat task_4.yaml
- hosts: webserver
  tasks:
    - name: Service httpd Restarted
      service: name={{ item }} state=restarted
     # with_items:
      loop:
        - httpd
        - memcached

ansible-playbook task_4.yaml -i hosts

例3 安装多个软件包

[root@db2 memcache]# cat task5.yaml
- hosts: webserver
  tasks:
    - name: Installed Httpd Memcached Package
      yum: name={{ package }} state=present
      vars:
        package:
          - httpd
          - memcached

ansible-playbook task5.yaml -i hosts

7、include和include_tasks

基本一样，include_tasks执行会显示调用的yaml的实际路径，而include是透明的

[root@db2 memcache]# cat restart_httpd.yaml
- name: Restart httpd server
  service: name=httpd state=restarted
  
[root@db2 memcache]# cat a_project.yaml
- hosts: webserver
  tasks:
    - name: A project command
      command: echo "A"

    - name: Restart httpd  and memcached
      include: restart_httpd.yaml
      
[root@db2 memcache]# cat b_project.yaml
- hosts: webserver
  tasks:
    - name: B project command
      command: echo "B"
    - name: Restart httpd memcached
      include_tasks: restart_httpd.yaml

8、import_playbook和import_tasks

import_playbook的剧本（文件）必须完整包括hosts和tasks

[root@db2 memcache]# cat import_playbook.yaml
- name: this is import_play
  hosts: webserver
  tasks:
    - name: install httpd
      yum: name=httpd state=latest
      
[root@db2 memcache]# cat playbook.yaml
- name: import
  import_playbook: import_playbook.yaml
- name: this is main_play
  hosts: webserver
  tasks:
    - name: restart httpd
      service: name=httpd state=restarted enabled=yes

import_tasks的剧本（文件）不需要hosts和tasks

​

[root@db2 memcache]# cat import_tasks.yaml
- name: install httpd
  yum: name=httpd state=latest
  
  [root@db2 memcache]# cat playbook.yaml
- name: this is main_play
  hosts: webserver
  tasks:
    - name: import
      import_tasks: import_tasks.yaml

9、changed_when

有些时候某些tasks只是去被控端获取一些数据，并不会对被控端做任何修改和变更，执行playbook的时候颜色还是黄色。

设置完后，再次执行，已经设置的tasks会显示绿色

tasks:
    - shell: uptime |awk \'{printf("%.2f",$(NF-2))}\'
      register: result
      changed_when: false

检查nginx -t配置文件状态,如果结果有successful，则执行的时候不显示发生改变，即颜色为绿色

- name: Check nginx configure status
  command: /usr/sbin/nginx -t
  register: check_nginx
  change_when:
    - ( check_nginx.stdout.find(\'successful\'))
    - false