ansible——ansible的变量

一、ansible变量的作用

所谓的变量就是一个会变的值。

如果你有一个值，被后面的task，频繁的调用，那么你就可以将这个值定义成一个变量，然后将后面task只用的这个值换成变量。比如有10处调用了。在playbook开头定义这个变量。当开头的变量值改变之后，10处被使用这个值的地方，就都做了改变。

Ansible支持变量功能，能将value存储到变量中，这样就能在Ansible项目中重复使用了。这样就可以简化项目的创建和维护，减少错误率 。

变量提供了一个变量提供了一种方便的方法来管理ansible项目中给定环境的动态值。 变量可以是要创建的用户，要安装的包，要启动的服务，要删除的文件，要从互联网下载的文件等等。

变量由必须以字母开头的字符串组成，并且只能包含字母，数字和下划线组成。

二、定义变量

变量几乎可以定义在ansible项目中的任意位置。然而这有三种基本的范围级别：

• Globe scope：从命令行设置变量或Ansible配置文件中设置变量

• play scope：在play和相关结构中设置变量

• host scope：通过inventory在主机组或单个主机中设置变量，fact采集变量或者register

如果在一个级别上定义了名字相同的两个变量，会选择先定义的那个。 在inventory中定义的变量会被playbook中定义的覆盖，playbook中定义的变量会被命令行定义的变量覆盖。

cmd > playbook > inventory

1、在playbook中定义变量

[root@control ansible_manage]# cat play_book/debug2.yaml 
---
- name: this is a var testhosts: allvars:user: nanatasks:- name: debug moudle textdebug:msg: "{{ user }} is very handsome man ,but {{user}} is  very pool"

vars这里可以定义变量的值，在这里定义的变量是"user=nana"

引用变量必须以"{{ vars }}"的格式来写

[root@control ansible_manage]# ansible-playbook play_book/debug2.yaml 
...
ok: [manage1] => {"msg": "nana is very handsome man ,but nana is  very pool"
}
ok: [manage3] => {"msg": "nana is very handsome man ,but nana is  very pool"
}
ok: [manage2] => {"msg": "nana is very handsome man ,but nana is  very pool"
}
...

可以看到输出的结果是变量引用的值

2、在cmd中定义变量

[root@control ansible_manage]# ansible-playbook play_book/debug2.yaml -e user=cui
...
ok: [manage1] => {"msg": "cui is very handsome man ,but cui is  very pool"
}
ok: [manage3] => {"msg": "cui is very handsome man ,but cui is  very pool"
}
ok: [manage2] => {"msg": "cui is very handsome man ,but cui is  very pool"
}
...

在yaml文件中的变量是uesr=nana，但是在cmd命令行中的定义了user=cui，这个时候变量就使用的是cmd中定义的变量，cmd中定义的变量优先级最高

3、在inventory中定义变量

单个设置

[root@control ansible_manage]# cat inventory 
manage1 user=jojo
manage2 user=dio
manage3 user=haha

[root@control ansible_manage]# ansible-playbook play_book/debug2.yaml
...
ok: [manage1] => {"msg": "nana is very handsome man ,but nana is  very pool"
}
ok: [manage3] => {"msg": "nana is very handsome man ,but nana is  very pool"
}
ok: [manage2] => {"msg": "nana is very handsome man ,but nana is  very pool"
...

由于在playbook中变量的优先级比较高，先将play中的变量删除

[root@control ansible_manage]# cat play_book/debug2.yaml 
---
- name: this is a var testhosts: allvars: tasks:- name: debug moudle textdebug:msg: "{{ user }} is very handsome man ,but {{user}} is  very pool"

再次执行

[root@control ansible_manage]# ansible-playbook play_book/debug2.yaml"msg": "jojo is very handsome man ,but jojo is  very pool"
}
ok: [manage2] => {"msg": "dio is very handsome man ,but dio is  very pool"
}
ok: [manage3] => {"msg": "haha is very handsome man ,but haha is  very pool"
}

使用组来设置

查看inventory文件

[root@control ansible_manage]# cat inventory 
manage1 
manage2
manage3 user=nana[computer]
manage1
manage2
manage3[computer:vars]         ## vars 表示定义这个组的变量，下面跟变量
user=mooooom

在这里定义了组的变量，computer组的变量user=mooooom，还定义了manage3用户的变量uers=nana

playbook文件

[root@control ansible_manage]# cat play_book/debug2.yaml
---
- name: this is a var testhosts: computervars:tasks:- name: debug moudle textdebug:msg: "{{ user }} is very handsome man ,but {{ user }} is  very pool"

执行文件

[root@control ansible_manage]# ansible-playbook play_book/debug2.yaml 
..."msg": "mooooom is very handsome man ,but mooooom is  very pool"
}
ok: [manage2] => {"msg": "mooooom is very handsome man ,but mooooom is  very pool"
}
ok: [manage3] => {"msg": "nana is very handsome man ,but nana is  very pool"
}
...

当inventory文件中有重复的变量出现的时候，有限范围小的变量

4、变量文件

将变量放在playbook和inventory中是非常不可取的一种方案，容易有很多的问题，管理也不方便，但是将变量单独放在一个文件中的话，就不会有这样的问题

inventory文件

[root@control ansible_manage]# cat inventory 
manage1 
manage2
manage3 user=nana[computer]
manage1
manage2
manage3

 playbook文件

[root@control ansible_manage]# cat play_book/debug2.yaml
---
- name: this is a var testhosts: allvars_files:- ./debug2_vars.yamltasks:- name: debug moudle textdebug:msg: "{{ user1 }} is very {{ state1 }} man ,but {{ user2 }} is  very {{ state2 }}"

变量文件

[root@control ansible_manage]# cat play_book/debug2_vars.yaml 
user1: nana
user2: cui
state1: handsome
state2: poor

执行结果

[root@control ansible_manage]# ansible-playbook play_book/debug2.yaml 
...
*******************************************************************************************ok: [manage1] => {"msg": "nana is very handsome man ,but cui is  very poor"
}
ok: [manage2] => {"msg": "nana is very handsome man ,but cui is  very poor"
}
ok: [manage3] => {"msg": "nana is very handsome man ,but cui is  very poor"
}

三、主机变量与主机组变量

playbook文件中放的就是task

inventory 文件中放的就是host and group

variables文件中放的就是var files

每个文件都放自己应该放的内容，如果将变量放进其他两个文件中，就会变得非常不好管理。

如何定义主机变量和主机组变量，在ansible中定义了两个文件，用来存放主机变量和组变量， group_vars和host_vars两个文件

• host_vars文件下面创建对应主机的文本，创建的文本名称名称是与inventory文件中host对应的文本

• gruop _vars文件下面创建对应主机组的文本，创建的文本名称必须是inventory文件中gruop对应的文本

1、inventory文件与host_vars和gruop _vars之间的关系

• inventory、host_vars、gruop _vars在同一目录下，或与对应playbook同于目录下

• host_vars文件的命名必须是"host_vars"，否则识别不到

• gruop _vars文件的命名必须是"gruop _vars"，否则识别不到

• host_vars下的文件名必须与inventory文件中的主机名对应

• gruop _vars下的文件必须与inventory文件中的主机组对应

[root@control ansible_manage]# cat inventory
manage1 
manage2
manage3 user=nana[computer]
manage1
manage2
manage3

[root@control ansible_manage]# tree
.
├── ansible.cfg
├── group_vars
│   └── computer
├── host_vars
│   ├── manage1
│   ├── manage2
│   └── manage3
├── inventory
└── play_book└── debug3.yaml

 比如在这个group_vars文件中的文件名就与inventory中的computer组的名字相同

2、实例

inventory文件

[root@control ansible_manage]# cat inventory
manage1 
manage2
manage3 user=nana[computer]
manage1
manage2
manage3

host_vars文件

[root@control ansible_manage]# cat host_vars/manage{1..3}
hello: wordssssssssssssssssssssssss    # manage1
hello: china                           # manage2
hello: ansible                         # manage3

group_vars文件

[root@control ansible_manage]# cat group_vars/computer 
user1: cui
state1: handsome

yaml文件

[root@control ansible_manage]# cat play_book/debug3.yaml 
---
- name: this is a var testhosts: alltasks:- name: debug moudle textdebug:var: user1    ## var后面可以直接跟变量，不需要"{{}}"- debug:var: state1- debug:var: hello- debug:var: nihao

执行文件

[root@control ansible_manage]# ansible-playbook play_book/debug3.yaml 
...
TASK [debug moudle text] *******************************************************************************************ok: [manage1] => {"user1": "cui"
}
ok: [manage2] => {"user1": "cui"
}
ok: [manage3] => {"user1": "cui"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"state1": "handsome"
}
ok: [manage2] => {"state1": "handsome"
}
ok: [manage3] => {"state1": "handsome"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"hello": "wordssssssssssssssssssssssss"
}
ok: [manage2] => {"hello": "china"
}
ok: [manage3] => {"hello": "ansible"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"nihao": "VARIABLE IS NOT DEFINED!"
}
ok: [manage2] => {"nihao": "VARIABLE IS NOT DEFINED!"
}
ok: [manage3] => {"nihao": "VARIABLE IS NOT DEFINED!"
}
...

四、变量矩阵

如果需要设置许多十分近似的变量，就可以使用变量矩阵的方式来写

例如：

user_nana_name=jinzhuren
user_nana_money=10000w
user_nana_home=sxpi
user_cui_name=cheng
user_cui_money=1k
user_cui_home=scdd

就可以写成这种模式

user:nana:name: jinzhurenmoney: 10000whome: sxpicui:name: chengmoney: 1khome: scdd

在playbook中定义变量

---
- name: this is a var testhosts: manage1vars_files:./debug4_var.yamltasks:- name: debug moudle textdebug:var: user['nana']['name']   ##  建议使用这种方式写变量，虽然这种写法真麻烦- debug:var: user['nana']['money']- debug:var: user['nana']['home']- debug:var: user.cui.name         ## 也可以使用这种方式写变量，但是容易与python冲突。- debug:var: user.cui.home- debug:var: user.cui.money

执行

[root@control ansible_manage]# ansible-playbook play_book/debug4.yaml
...
TASK [debug moudle text] *******************************************************************************************ok: [manage1] => {"user['nana']['name']": "jinzhuren"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"user['nana']['money']": "10000w"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"user['nana']['home']": "sxpi"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"user.cui.name": "cheng"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"user.cui.home": "scdd"
}TASK [debug] *******************************************************************************************************ok: [manage1] => {"user.cui.money": "1k"
}
...

五、register变量

可以将tasks运行状态记录下来

写一个install.yaml

[root@control ansible_manage]# cat play_book/install.yaml
---
- name: install a pkghosts: manage1tasks:- name: install httpdyum:name: httpdstate: presentregister: ergou      ## 使用ergou这个变量，记录运行状态- name: degue ergoudebug:var: ergou         ## 条用这个二狗

[root@control ansible_manage]# ansible-playbook play_book/install.yaml
...
TASK [degue ergou] *****************************************************************************************************************************ok: [manage1] => {"ergou": {"changed": true,"failed": false,"msg": "","rc": 0,"results": ["Installed: mod_http2-1.15.19-5.el9.x86_64","Installed: httpd-2.4.57-5.el9.x86_64"]}
}PLAY RECAP *************************************************************************************************************************************manage1                    : ok=3    changed=1    unreachable=0    failed=0    skipped=0    rescued=0    ignored=0

运行之后可以看到使用register变量，搭配debug显示出了运行状态

rc的值为0，表示运行正常，若不为0，则表示出现了问题。常常使用rc的返回值来做条件判断

六、facts变量

能够轻松的获取到被控制节点主机的信息

获取其他host节点的信息

[root@control ansible_manage]# ansible manage1 -m setup

生成的是一个json文件，它里面包含了被管理主机的绝大多数信息，可以很轻易的知到被管理主机的资源信息等。它使用的是ansible的setup模块

在每次运行playbook的时候，ansible会默认的自动运行这个模块

1、使用debug模块获取主机的内存

使用ansible manage1 -m setup获取到合适的变量，进行筛选，编写playbook

[root@control ansible_manage]# cat play_book/debug5.yaml 
---
- name: this is a var testhosts: alltasks:- name: debug moudle textdebug:msg: "{{ ansible_hostname }}'s current free mem is {{ ansible_memfree_mb }}"

[root@control ansible_manage]# ansible-playbook play_book/debug5.yaml PLAY [this is a var test] ***************************************************************************************## [Gathering Facts] 这个就是自动执行了setup模块获取信息
TASK [Gathering Facts] ******************************************************************************************
ok: [manage1]
ok: [manage3]
ok: [manage2]TASK [debug moudle text] ****************************************************************************************
ok: [manage1] => {"msg": "manage1's current free mem is 69"
}
ok: [manage2] => {"msg": "manage2's current free mem is 67"
}
ok: [manage3] => {"msg": "manage3's current free mem is 872"
}
...

运行结束后，可以看到返回的信息，获取到了三台主机的内存信息

2、关闭facts变量的采集

如果集群过于大的情况下，采集信息是非常消耗时间的一件事情。

在playbook中使用gather_facts: no就可以变比变量的采集

[root@control ansible_manage]# cat play_book/debug6.yaml
---
- name: this is a var testhosts: allgather_facts: notasks:- name: debug moudle textdebug:var:- ansible_hostname- ansible_memfree_mb

[root@control ansible_manage]# ansible-playbook play_book/debug6.yamlPLAY [this is a var test] ***************************************************************************************TASK [debug moudle text] ****************************************************************************************
ok: [manage1] => {"<class 'list'>": "VARIABLE IS NOT DEFINED!"
}
ok: [manage2] => {"<class 'list'>": "VARIABLE IS NOT DEFINED!"
}
ok: [manage3] => {"<class 'list'>": "VARIABLE IS NOT DEFINED!"
}
...

可以看到已经没有使用setbug模块

3、自定义faces的变量

在faces中有一个"ansible_local": {},的变量，这个变量是提供给用户自定义变量使用的

 [root@control ansible_manage]# ansible manage1 -m setup
..."ansible_local": {},
...

自定义变量默认的目录是/etc/ansible/facts.d/目录，创建的文件必须是以.fact结尾，这个目录如果没有，就需要自己创建（必须要在被管理主机上创建，必须有root的权限）

定义一个文件

[root@manage1 ~]# mkdir -p /etc/ansible/facts.d/
[root@manage1 ~]# vim /etc/ansible/facts.d/test.fact
[root@manage1 ~]# cat /etc/ansible/facts.d/test.fact 
[nana]
name=nananananananananana

[root@control ansible_manage]# ansible manage1 -m setup | grep -A 5 ansible_local"ansible_local": {"test": {"nana": {"name": "nananananananananana"}}

可以看到ansible_local已经出现了值

playbook使用debug调用一下

[root@control ansible_manage]# cat play_book/debug7.yaml 
---
- name: this is a var testhosts: manage1tasks:- name: debug moudle textdebug:var: ansible_local['test']['nana']['name']

[root@control ansible_manage]# ansible-playbook play_book/debug7.yamlPLAY [this is a var test] ***************************************************************************************TASK [Gathering Facts] ******************************************************************************************
ok: [manage1]TASK [debug moudle text] ****************************************************************************************
ok: [manage1] => {"ansible_local['test']['nana']['name']": "nananananananananana"
}PLAY RECAP ******************************************************************************************************
manage1                    : ok=2    changed=0    unreachable=0    failed=0    skipped=0    rescued=0    ignored=
0

七、magic变量

就是主机变量和主机组变量

[root@control ansible_manage]# ansible manage1 -m debug -a "var=hostvars['manage1']"
manage1 | SUCCESS => {"hostvars['manage1']": {"ansible_check_mode": false,"ansible_config_file": "/root/ansible_manage/ansible.cfg","ansible_diff_mode": false,"ansible_facts": {},"ansible_forks": 5,"ansible_inventory_sources": ["/root/ansible_manage/inventory"],"ansible_playbook_python": "/usr/bin/python3","ansible_verbosity": 0,"ansible_version": {"full": "2.14.9","major": 2,"minor": 14,"revision": 9,"string": "2.14.9"},"group_names": ["computer"],"groups": {"all": ["manage1","manage2","manage3"],"computer": ["manage1","manage2","manage3"],"ungrouped": []},"hello": "wordssssssssssssssssssssssss","inventory_dir": "/root/ansible_manage","inventory_file": "/root/ansible_manage/inventory","inventory_hostname": "manage1","inventory_hostname_short": "manage1","omit": "__omit_place_holder__1d6935aaefc1e64ad57756c57c90f8ad4cc240af","playbook_dir": "/root/ansible_manage","state1": "handsome","user1": "cui"}
}

最有用的是inventory_hostname和groups，它可以将用于条件判断