ansible系列(28)--ansible的playbook异常处理

1. playbook的异常处理
- 1.1 Playbook错误忽略
- 1.2 task执行失败强制调用handlers
1.3 控制Tasks报告状态为OK
- 1.4 changed_when检查任务结果

1. playbook的异常处理

1.1 Playbook错误忽略

在 playbook 执行的过程中，难免会遇到一些错误（返回结果不为0）。由于 playbook 遇到错误后，不会执行之后的任务，不便于调试，此时，可以使用 ignore_errors 来暂时忽略错误，使得 playbook 继续执行。

首先编写一个playbook文件：

[root@xuzhichao playbook]# cat ignore.yml

- hosts: localhost

  tasks:

    - name: Igonre False

      shell: /bin/false    <==此命令返回值一定不为0

    - name: Print Message

      debug:

        msg: "It's hard to here"

此时运行playbook时会在Igonre False任务中报错，无法继续向下运行：

[root@xuzhichao playbook]# ansible-playbook ignore.yml 

PLAY [localhost] **********************************************************************************************************************************************

TASK [Igonre False] *******************************************************************************************************************************************

fatal: [localhost]: FAILED! => {"changed": true, "cmd": "/bin/false", "delta": "0:00:00.024018", "end": "2021-08-05 21:36:36.638534", "msg": "non-zero return code", "rc": 1, "start": "2021-08-05 21:36:36.614516", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}

PLAY RECAP ****************************************************************************************************************************************************

localhost                  : ok=0    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0

修改playbook文件，忽略中间task的错误，继续向下执行：

[root@xuzhichao playbook]# cat ignore.yml

- hosts: localhost

  tasks:

    - name: Igonre False

      shell: /bin/false

      ignore_errors: yes

    - name: Print Message

      debug:

        msg: "It's hard to here"

此时运行playbook会忽略掉中间task的错误，继续向下运行：

[root@xuzhichao playbook]# ansible-playbook ignore.yml 

PLAY [localhost] **********************************************************************************************************************************************

TASK [Igonre False] *******************************************************************************************************************************************

fatal: [localhost]: FAILED! => {"changed": true, "cmd": "/bin/false", "delta": "0:00:00.004268", "end": "2021-08-05 21:37:16.630140", "msg": "non-zero return code", "rc": 1, "start": "2021-08-05 21:37:16.625872", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}

...ignoring      <==忽略错误

TASK [Print Message] ******************************************************************************************************************************************

ok: [localhost] => {

    "msg": "It's hard to here"

}

PLAY RECAP ****************************************************************************************************************************************************

localhost                  : ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=1

也可以写成下面的形式来保证task的返回值为0：

- hosts: localhost

  tasks:

    - name: Igonre False

      shell: /bin/false || /bin/ture

    - name: Print Message

      debug:

        msg: "It's hard to here"

1.2 task执行失败强制调用handlers

通常情况下，当 task 失败后， playbook 将会终止，任何在前面已经被 tasks notify 的 handlers 都不会被执行。如果你在 playbook 中设置了 force_handlers: yes 参数，被通知的 handlers 就会被强制执行。

请看下面的playbook：

[root@xuzhichao playbook]# cat force_handlers.yml

- hosts: localhost

  tasks:

    - name: Copy Configure File

      template:

        src: conf/nginx.conf.j2

        dest: /tmp/nginx.conf

      notify: Start Nginx

    - name: Igonre False

      shell: /bin/false

  handlers:

    - name: Start Nginx

      service:

        name: nginx

        state: started

当conf/nginx.conf.j2文件发生改变时，由于在 Igonre False这个task中执行出现错误，将不会执行handlers中的内容。

[root@xuzhichao playbook]# ansible-playbook force_handlers.yml 

PLAY [localhost] **********************************************************************************************************************************************

TASK [Copy Configure File] ************************************************************************************************************************************

ok: [localhost]

TASK [Igonre False] *******************************************************************************************************************************************

fatal: [localhost]: FAILED! => {"changed": true, "cmd": "/bin/false", "delta": "0:00:00.006289", "end": "2021-08-05 22:09:40.799988", "msg": "non-zero return code", "rc": 1, "start": "2021-08-05 22:09:40.793699", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}

PLAY RECAP ****************************************************************************************************************************************************

localhost                  : ok=1    changed=0    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0

若此时需要强制执行handlers中的内容，需要使用force_handlers: yes 参数，playbook修改如下：

[root@xuzhichao playbook]# cat force_handlers.yml

- hosts: localhost

  force_handlers: yes

  tasks:

    - name: Copy Configure File

      template:

        src: conf/nginx.conf.j2

        dest: /tmp/nginx.conf

      notify: Start Nginx

    - name: Igonre False

      shell: /bin/false

  handlers:

    - name: Start Nginx

      service:

        name: nginx

        state: started

此时运行playbook，当conf/nginx.conf.j2文件发生改变时，handlers将会被强制执行：

[root@xuzhichao playbook]# ansible-playbook force_handlers.yml 

PLAY [localhost] **********************************************************************************************************************************************

TASK [Copy Configure File] ************************************************************************************************************************************

changed: [localhost]

TASK [Igonre False] *******************************************************************************************************************************************

fatal: [localhost]: FAILED! => {"changed": true, "cmd": "/bin/false", "delta": "0:00:00.003843", "end": "2021-08-05 22:12:00.335807", "msg": "non-zero return code", "rc": 1, "start": "2021-08-05 22:12:00.331964", "stderr": "", "stderr_lines": [], "stdout": "", "stdout_lines": []}

RUNNING HANDLER [Start Nginx] *********************************************************************************************************************************

ok: [localhost]

PLAY RECAP ****************************************************************************************************************************************************

localhost                  : ok=2    changed=1    unreachable=0    failed=1    skipped=0    rescued=0    ignored=0

1.3 控制Tasks报告状态为OK

有时一些task会在每次playbook运行是都会执行一次，这样会导致playbook的运行结果中该task会显示为changed状态，但是有时这些task并没有改变被控端的状态，仅仅是获取了某些信息，此时可以使用changed_when: false 来抑制这个状态。

请看下面示例：

playbook文件内容如下：

[root@xuzhichao playbook]# cat status_ok.yml

- hosts: localhost

  tasks:

    - name: Get Nginx Server Port

      shell: netstat -ntlp | grep nginx

      register: nginx_port

    - name: Print Nginx Port

      debug:

        msg:

          - "{{ nginx_port.stdout_lines }}"

每次的执行结果“Get Nginx Server Port”任务状态都是changed：

[root@xuzhichao playbook]# ansible-playbook status_ok.yml 

PLAY [localhost] **********************************************************************************************************************************************

TASK [Get Nginx Server Port] **********************************************************************************************************************************

changed: [localhost]     <==状态为changed

TASK [Print Nginx Port] ***************************************************************************************************************************************

ok: [localhost] => {

    "msg": [

        [

            "tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      26579/nginx: master "

        ]

    ]

}

PLAY RECAP ****************************************************************************************************************************************************

localhost                  : ok=2    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

可以使用changed_when: false 来抑制这个状态，playbook修改如下：

[root@xuzhichao playbook]# cat status_ok.yml

- hosts: localhost

  tasks:

    - name: Get Nginx Server Port

      shell: netstat -ntlp | grep nginx

      register: nginx_port

      changed_when: false     <==该task不再发生changed提示

    - name: Print Nginx Port

      debug:

        msg:

          - "{{ nginx_port.stdout_lines }}"

此时执行结果如下：

[root@xuzhichao playbook]# ansible-playbook status_ok.yml 

PLAY [localhost] **********************************************************************************************************************************************

TASK [Get Nginx Server Port] **********************************************************************************************************************************

ok: [localhost]     <==变为OK状态

TASK [Print Nginx Port] ***************************************************************************************************************************************

ok: [localhost] => {

    "msg": [

        [

            "tcp        0      0 0.0.0.0:80              0.0.0.0:*               LISTEN      26579/nginx: master "

        ]

    ]

}

PLAY RECAP ****************************************************************************************************************************************************

localhost                  : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

1.4 changed_when检查任务结果

仍然以安装httpd为例，在每次修改完httpd的配置文件后，都需要检查配置文件的语法是否正确，如果不正确，则退出后续task，不重启httpd服务。

playbook文件优化如下：

[root@xuzhichao playbook]# cat change_when.yml

---

- hosts: 192.168.20.23

  remote_user: root

  tasks:

    - name: Install Htttpd Server

      yum:

        name: httpd

        state: present

      tags:

        - install_httpd

        - install_apache

    - name: Configure Httpd Server

      copy:

        src: conf/httpd.conf.j2

        dest: /etc/httpd/conf/httpd.conf

        owner: "root"

        group: "root"

        mode: "0644"

      notify: Restart Httpd Server

      tags: conf_httpd

    - name: Check Httpd Configuer        <==检测配置文件语法

      shell: /usr/sbin/httpd -t

      register: httpd_check

      changed_when:

        - httpd_check.stdout.find('OK')  <==查找变量返回的结果是否有ok，如不存在则终止该tasks

        - false     <==被管理主机没有发生变化状态则为OK

    - name: Init Httpd Server

      copy:

        src: file/test.html.j2

        dest: /var/www/html/test.html

        owner: "apache"

        group: "apache"

        mode: "0644"

      tags: init_httpd

    - name: Start Httpd Server

      service:

        name: httpd

        state: started

        enabled: yes

      tags: start_httpd

  handlers:

    - name: Restart Httpd Server

      service:

        name: httpd

        state: restarted

      tags: restart httpd

[root@nginx03 ~]# httpd -t

Syntax OK

playbook运行结果如下：

[root@xuzhichao playbook]# ansible-playbook change_when.yml 

PLAY [192.168.20.23] ******************************************************************************************************************************************

TASK [Install Htttpd Server] **********************************************************************************************************************************

ok: [192.168.20.23]

TASK [Configure Httpd Server] *********************************************************************************************************************************

ok: [192.168.20.23]

TASK [Check Httpd Configuer] **********************************************************************************************************************************

ok: [192.168.20.23]

TASK [Init Httpd Server] **************************************************************************************************************************************

ok: [192.168.20.23]

TASK [Start Httpd Server] *************************************************************************************************************************************

changed: [192.168.20.23]

PLAY RECAP ****************************************************************************************************************************************************

192.168.20.23              : ok=5    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0