今天碰到一个需求，当使用ansible-playbook在执行一个脚本后，根据脚本返回的内容判断是否继续往下执行还是中断执行，查询官网发现使用register寄存器可以实现记录脚本输出，使用when+fail模块来判断是否往下继续执行或者中断，以下是一个简单例子：

ansible服务在172.16.133.128上，先执行一个脚本，脚本返回success时在172.16.133.129上创建一个目录/tmp/test，如果返回failed则中断执行

先在172.16.133.129上配置一个简单脚本/tmp/test.sh

#!/bin/bash
echo "failed"

配置yml文件test.yml，先配置执行脚本部分

• hosts: 172.16.133.129
  remote_user: root
  tasks:
  • command: /tmp/test.sh
    register: result
    这里使用了register寄存器，具体寄存了什么内容，可以使用-v参数来查看输出

使用ansible-playbook运行

[root@localhost ansible]# ansible-playbook -v test.yml
PLAY [172.16.133.129] *
GATHERING FACTS ***
ok: [172.16.133.129]
TASK: [command /tmp/test.sh] **
changed: [172.16.133.129] => {"changed": true, "cmd": ["/tmp/test.sh"], "delta": "0:00:00.002602", "end": "2016-04-11 17:00:57.227517", "rc": 0, "start": "2016-04-11 17:00:57.224915", "stderr": "", "stdout": "failed", "warnings": []}
PLAY RECAP ****
172.16.133.129 : ok=2 changed=1 unreachable=0 failed=0
register保存的信息就时“=>”后面的字典信息，信息保存在result变量中

可以看到"stdout"就是脚本的标准输出信息，这时可以使用"when"来判断是否执行或者跳过

修改test.yml

• hosts: 172.16.133.129
  remote_user: root
  tasks:

  • command: /tmp/test.sh
    register: result

  • file: path=/tmp/test state=directory
    when: result.stdout == ‘failed‘
    当脚本输出为‘failed‘时就会执行创建目录操作，输出不为‘failed‘，则跳过执行

执行yml后输出如下：

[root@localhost ansible]# ansible-playbook -v test.yml

PLAY [172.16.133.129] *****

GATHERING FACTS ***
ok: [172.16.133.129]

TASK: [command /tmp/test.sh] **
changed: [172.16.133.129] => {"changed": true, "cmd": ["/tmp/test.sh"], "delta": "0:00:00.002618", "end": "2016-04-11 17:11:07.957097", "rc": 0, "start": "2016-04-11 17:11:07.954479", "stderr": "", "stdout": "failed", "warnings": []}

TASK: [file path=/tmp/test state=directory] ***
changed: [172.16.133.129] => {"changed": true, "gid": 0, "group": "root", "mode": "0755", "owner": "root", "path": "/tmp/test", "size": 4096, "state": "directory", "uid": 0}

PLAY RECAP ****
172.16.133.129 : ok=3 changed=2 unreachable=0 failed=0

修改test.sh，输出‘success‘，然后执行test.yml

[root@localhost ansible]# ansible-playbook -v test.yml

PLAY [172.16.133.129] *****

GATHERING FACTS ***
ok: [172.16.133.129]

TASK: [command /tmp/test.sh] **
changed: [172.16.133.129] => {"changed": true, "cmd": ["/tmp/test.sh"], "delta": "0:00:00.002611", "end": "2016-04-11 17:14:08.456293", "rc": 0, "start": "2016-04-11 17:14:08.453682", "stderr": "", "stdout": "success", "warnings": []}

TASK: [file path=/tmp/test state=directory] ***
skipping: [172.16.133.129]

PLAY RECAP ****
172.16.133.129 : ok=2 changed=1 unreachable=0 failed=0
可以看到当stdout为success时，创建目录的操作直接跳过了

现在回到最终要达到的目的，当脚本输出为‘failed‘时，我们要中断ansible-playbook执行，而不是跳过，这时需要配合fail模块来实现，并且可以抛出自定义信息

修改test.sh文件,输出‘failed‘

修改test.yml

• hosts: 172.16.133.129
  remote_user: root
  tasks:

  • command: /tmp/test.sh
    register: result

  • name: if stdout ‘failed‘ ,Interrupt execution
    fail: msg="check failed"
    when: result.stdout == ‘failed‘

  • file: path=/tmp/test state=directory
    如果脚本执行返回‘failed‘，中断执行，并抛出信息"check failed"

[root@localhost ansible]# ansible-playbook -v test.yml

PLAY [172.16.133.129] *****

GATHERING FACTS ***
ok: [172.16.133.129]

TASK: [command /tmp/test.sh] **
changed: [172.16.133.129] => {"changed": true, "cmd": ["/tmp/test.sh"], "delta": "0:00:00.002632", "end": "2016-04-11 17:27:47.248996", "rc": 0, "start": "2016-04-11 17:27:47.246364", "stderr": "", "stdout": "failed", "warnings": []}

TASK: [check ,Interrupt execution] ****
failed: [172.16.133.129] => {"failed": true}
msg: check failed

FATAL: all hosts have already failed -- aborting

PLAY RECAP ****
to retry, use: --limit @/root/test.retry

172.16.133.129 : ok=2 changed=1 unreachable=0 failed=1
playbook运行到第二个task时就会报错并抛出msg

ansible官网文档：http://docs.ansible.com/ansible/playbooks_conditionals.html#register-variables

注：寄存器中的数据都可以拿来使用，如"rc","stderr"等，当然也有很多种方法，文中的‘failed‘是严格匹配，也可以使用模糊查找，如result.stdout.find(‘failed‘) != -1也可以达到相同的效果

本文出自 “漂泊的鱼” 博客，请务必保留此出处http://faded.blog.51cto.com/6375932/1762688