Ansible playbook处理程序不通过notify语句启动服务

Question

我不明白或看到我做错了什么,但似乎我无法让我的ansible playbook在ubuntu服务器上启动rng-tools服务.

environment:
    remote server (configured): Ubuntu 14.04.1 LTS
    server with playbook: Debian GNU/Linux 7.6 (wheezy)
    > apt-cache policy ansible
    ansible:
      Installed: 1.7-0.git201406241728~unstable
      Candidate: 1.7-0.git201406241728~unstable
      Version table:
      *** 1.7-0.git201406241728~unstable 0
            100 /var/lib/dpkg/status

我的文件夹结构如下:

tasks/main.yml
tasks/packages.yml that is included from the main.yml file
handlers/main.yml
files/rng-tools    

我正在使用我的playbook安装包rng-tools,并在执行后安装.

- name: install common packages
  apt: name={{ item }}  state=present
  with_items:
    - gnupg
    - rng-tools
    - reprepro
    - dpkg-sig
    - nginx

是的我正在安装多个包,但它们都已正确安装.

安装完成后,我使用以下方法复制/ etc/default/rng-tools文件:

- name: copy urandom default conf file
  copy: src="rng-tools" dest="/etc/default/rng-tools"
  notify: start rng-tools 

您可以注意到我写了notify: restart rng-tools那里以启动服务.

我的handlers/main.yml文件看起来像:

---
# Handler for rng-tools
- name: start rng-tools
  service: name=rng-tools state=started

- name: restart rng-tools
  service: name=rng-tools state=restarted

- name: stop rng-tools
  service: name=rng-tools state=stopped

- name: reload rng-tools
  service: name=rng-tools state=reloaded

正在运行的playbook的日志产生以下内容:

TASK: [debian-repository | install common packages] ***************************
ok: [debian.home.dr] => (item=gnupg,rng-tools,reprepro,dpkg-sig,nginx) => {"changed": false, "item": "gnupg,rng-tools,reprepro,dpkg-sig,nginx"}

TASK: [debian-repository | copy urandom default conf file] ********************
ok: [debian.home.dr] => {"changed": false, "dest": "/etc/default/rng-tools", "gid": 0, "group": "root", "md5sum": "45ed1b1ee174494442296fdd262f3b09", "mode": "0644", "owner": "root", "path": "/etc/default/rng-tools", "size": 815, "state": "file", "uid": 0}

PLAY RECAP ********************************************************************
debian.home.dr             : ok=12   changed=2    unreachable=0    failed=0

问题是服务永远不会启动,我检查一下

sudo ps aux | grep rng
root 29392  0.0  0.0  11740   884 pts/1    S+   22:45   0:00 grep rng

当我这样做sudo service rng-tools start时,我可以看到这个过程运行正常.

~:sudo service rng-tools start
Starting Hardware RNG entropy gatherer daemon: rngd.
~:sudo ps aux | grep rng
root     29431  0.0  0.0   8964   336 ?        Ss   22:47   0:00 /usr/sbin/rngd -r /dev/urandom
root 29433  0.0  0.0  11740   880 pts/1    S+   22:47   0:00 grep rng

我尝试了很多东西,比如评论除了一个之外的所有处理程序.如果我在复制文件后输入命令类型并且命令启动服务运行服务,那么它似乎不是一个权限问题,但不知何故我无法弄明白.另外使用handlers/main.yml是最佳实践指南,所以我相信我做得对.

你知道为什么我不能开始使用剧本吗？

Answer 1

changed:false 是解开这个谜团的关键:

TASK: [debian-repository | copy urandom default conf file] ********************
ok: [debian.home.dr] => {"changed": false, "dest": "/etc/default/rng-tools", "gid": 0, "group": "root", "md5sum": "45ed1b1ee174494442296fdd262f3b09", "mode": "0644", "owner": "root", "path": "/etc/default/rng-tools", "size": 815, "state": "file", "uid": 0}

只有在发生文件复制时才会调用"notify:started".