原生的ansible不支持service.status,在Google之后,发现有人提交了一个patch,可以支持status选项.见https://github.com/ritzk/ansible-modules-core/commit/ab4c9c226e94371c2b60aab57ef5b0a76f75a74b

就此使用一段时间后,在centos 7使用时,出现了报错.

root@HOSTXXXXXXX001:~# ansible HOSTXXXXXXX003 -m service -a 'name="zabbix-agent" state="status"' HOSTXXXXXXX003 | FAILED >> {
"failed": true,
"msg": "value of state must be one of: running,started,stopped,restarted,reloaded, got: status"
}

以前都是在centos 6使用的,正常,如今在centos 7出现报错,就打开源码查看了一下.

发现centos 6是/usr/local/lib/python3.6/dist-packages/ansible/modules/system/service.py

而centos 7使用的则是/usr/local/lib/python3.6/dist-packages/ansible/modules/system/systemd.py , 该文件并没有进行patch

于是对该文件进行了修改,如下.

#!/usr/bin/python
# -*- coding: utf-8 -*- # Copyright: (c) 2016, Brian Coca <bcoca@ansible.com>
# GNU General Public License v3.0+ (see COPYING or https://www.gnu.org/licenses/gpl-3.0.txt)
# patch by Redheat
# email qjyyn@qq.com from __future__ import absolute_import, division, print_function __metaclass__ = type ANSIBLE_METADATA = {'metadata_version': '1.1',
'status': ['stableinterface'],
'supported_by': 'core'} DOCUMENTATION = '''
module: systemd
author:
- Ansible Core Team
version_added: "2.2"
short_description: Manage services
description:
- Controls systemd services on remote hosts.
options:
name:
description:
- Name of the service. When using in a chroot environment you always need to specify the full name i.e. (crond.service).
aliases: [ service, unit ]
state:
description:
- C(started)/C(stopped) are idempotent actions that will not run commands unless necessary.
C(restarted) will always bounce the service. C(reloaded) will always reload.
choices: [ reloaded, restarted, started, stopped ]
enabled:
description:
- Whether the service should start on boot. B(At least one of state and enabled are required.)
type: bool
masked:
description:
- Whether the unit should be masked or not, a masked unit is impossible to start.
type: bool
daemon_reload:
description:
- run daemon-reload before doing any other operations, to make sure systemd has read any changes.
type: bool
default: 'no'
aliases: [ daemon-reload ]
user:
description:
- run systemctl talking to the service manager of the calling user, rather than the service manager
of the system.
type: bool
default: 'no'
no_block:
description:
- Do not synchronously wait for the requested operation to finish.
Enqueued job will continue without Ansible blocking on its completion.
type: bool
default: 'no'
version_added: "2.3"
notes:
- Since 2.4, one of the following options is required 'state', 'enabled', 'masked', 'daemon_reload', and all except 'daemon_reload' also require 'name'.
- Before 2.4 you always required 'name'.
requirements:
- A system managed by systemd.
''' EXAMPLES = '''
- name: Make sure a service is running
systemd: state=started name=httpd
- name: stop service cron on debian, if running
systemd: name=cron state=stopped
- name: restart service cron on centos, in all cases, also issue daemon-reload to pick up config changes
systemd:
state: restarted
daemon_reload: yes
name: crond
- name: reload service httpd, in all cases
systemd:
name: httpd
state: reloaded
- name: enable service httpd and ensure it is not masked
systemd:
name: httpd
enabled: yes
masked: no
- name: enable a timer for dnf-automatic
systemd:
name: dnf-automatic.timer
state: started
enabled: True
- name: just force systemd to reread configs (2.4 and above)
systemd: daemon_reload=yes
''' RETURN = '''
status:
description: A dictionary with the key=value pairs returned from `systemctl show`
returned: success
type: complex
contains: {
"ActiveEnterTimestamp": "Sun 2016-05-15 18:28:49 EDT",
"ActiveEnterTimestampMonotonic": "8135942",
"ActiveExitTimestampMonotonic": "0",
"ActiveState": "active",
"After": "auditd.service systemd-user-sessions.service time-sync.target systemd-journald.socket basic.target system.slice",
"AllowIsolate": "no",
"Before": "shutdown.target multi-user.target",
"BlockIOAccounting": "no",
"BlockIOWeight": "1000",
"CPUAccounting": "no",
"CPUSchedulingPolicy": "0",
"CPUSchedulingPriority": "0",
"CPUSchedulingResetOnFork": "no",
"CPUShares": "1024",
"CanIsolate": "no",
"CanReload": "yes",
"CanStart": "yes",
"CanStop": "yes",
"CapabilityBoundingSet": "18446744073709551615",
"ConditionResult": "yes",
"ConditionTimestamp": "Sun 2016-05-15 18:28:49 EDT",
"ConditionTimestampMonotonic": "7902742",
"Conflicts": "shutdown.target",
"ControlGroup": "/system.slice/crond.service",
"ControlPID": "0",
"DefaultDependencies": "yes",
"Delegate": "no",
"Description": "Command Scheduler",
"DevicePolicy": "auto",
"EnvironmentFile": "/etc/sysconfig/crond (ignore_errors=no)",
"ExecMainCode": "0",
"ExecMainExitTimestampMonotonic": "0",
"ExecMainPID": "595",
"ExecMainStartTimestamp": "Sun 2016-05-15 18:28:49 EDT",
"ExecMainStartTimestampMonotonic": "8134990",
"ExecMainStatus": "0",
"ExecReload": "{ path=/bin/kill ; argv[]=/bin/kill -HUP $MAINPID ; ignore_errors=no ; start_time=[n/a] ; stop_time=[n/a] ; pid=0 ; code=(null) ; status=0/0 }",
"ExecStart": "{ path=/usr/sbin/crond ; argv[]=/usr/sbin/crond -n $CRONDARGS ; ignore_errors=no ; start_time=[n/a] ; stop_time=[n/a] ; pid=0 ; code=(null) ; status=0/0 }",
"FragmentPath": "/usr/lib/systemd/system/crond.service",
"GuessMainPID": "yes",
"IOScheduling": "0",
"Id": "crond.service",
"IgnoreOnIsolate": "no",
"IgnoreOnSnapshot": "no",
"IgnoreSIGPIPE": "yes",
"InactiveEnterTimestampMonotonic": "0",
"InactiveExitTimestamp": "Sun 2016-05-15 18:28:49 EDT",
"InactiveExitTimestampMonotonic": "8135942",
"JobTimeoutUSec": "0",
"KillMode": "process",
"KillSignal": "15",
"LimitAS": "18446744073709551615",
"LimitCORE": "18446744073709551615",
"LimitCPU": "18446744073709551615",
"LimitDATA": "18446744073709551615",
"LimitFSIZE": "18446744073709551615",
"LimitLOCKS": "18446744073709551615",
"LimitMEMLOCK": "65536",
"LimitMSGQUEUE": "819200",
"LimitNICE": "0",
"LimitNOFILE": "4096",
"LimitNPROC": "3902",
"LimitRSS": "18446744073709551615",
"LimitRTPRIO": "0",
"LimitRTTIME": "18446744073709551615",
"LimitSIGPENDING": "3902",
"LimitSTACK": "18446744073709551615",
"LoadState": "loaded",
"MainPID": "595",
"MemoryAccounting": "no",
"MemoryLimit": "18446744073709551615",
"MountFlags": "0",
"Names": "crond.service",
"NeedDaemonReload": "no",
"Nice": "0",
"NoNewPrivileges": "no",
"NonBlocking": "no",
"NotifyAccess": "none",
"OOMScoreAdjust": "0",
"OnFailureIsolate": "no",
"PermissionsStartOnly": "no",
"PrivateNetwork": "no",
"PrivateTmp": "no",
"RefuseManualStart": "no",
"RefuseManualStop": "no",
"RemainAfterExit": "no",
"Requires": "basic.target",
"Restart": "no",
"RestartUSec": "100ms",
"Result": "success",
"RootDirectoryStartOnly": "no",
"SameProcessGroup": "no",
"SecureBits": "0",
"SendSIGHUP": "no",
"SendSIGKILL": "yes",
"Slice": "system.slice",
"StandardError": "inherit",
"StandardInput": "null",
"StandardOutput": "journal",
"StartLimitAction": "none",
"StartLimitBurst": "5",
"StartLimitInterval": "10000000",
"StatusErrno": "0",
"StopWhenUnneeded": "no",
"SubState": "running",
"SyslogLevelPrefix": "yes",
"SyslogPriority": "30",
"TTYReset": "no",
"TTYVHangup": "no",
"TTYVTDisallocate": "no",
"TimeoutStartUSec": "1min 30s",
"TimeoutStopUSec": "1min 30s",
"TimerSlackNSec": "50000",
"Transient": "no",
"Type": "simple",
"UMask": "0022",
"UnitFileState": "enabled",
"WantedBy": "multi-user.target",
"Wants": "system.slice",
"WatchdogTimestampMonotonic": "0",
"WatchdogUSec": "0",
}
''' # NOQA from ansible.module_utils.basic import AnsibleModule
from ansible.module_utils.service import sysv_exists, sysv_is_enabled, fail_if_missing
from ansible.module_utils._text import to_native def is_running_service(service_status):
return service_status['ActiveState'] in set(['active', 'activating']) def request_was_ignored(out):
return '=' not in out and 'ignoring request' in out def parse_systemctl_show(lines):
# The output of 'systemctl show' can contain values that span multiple lines. At first glance it
# appears that such values are always surrounded by {}, so the previous version of this code
# assumed that any value starting with { was a multi-line value; it would then consume lines
# until it saw a line that ended with }. However, it is possible to have a single-line value
# that starts with { but does not end with } (this could happen in the value for Description=,
# for example), and the previous version of this code would then consume all remaining lines as
# part of that value. Cryptically, this would lead to Ansible reporting that the service file
# couldn't be found.
#
# To avoid this issue, the following code only accepts multi-line values for keys whose names
# start with Exec (e.g., ExecStart=), since these are the only keys whose values are known to
# span multiple lines.
parsed = {}
multival = []
k = None
for line in lines:
if k is None:
if '=' in line:
k, v = line.split('=', 1)
if k.startswith('Exec') and v.lstrip().startswith('{'):
if not v.rstrip().endswith('}'):
multival.append(v)
continue
parsed[k] = v.strip()
k = None
else:
multival.append(line)
if line.rstrip().endswith('}'):
parsed[k] = '\n'.join(multival).strip()
multival = []
k = None
return parsed # ===========================================
# Main control flow def main():
# initialize
module = AnsibleModule(
argument_spec=dict(
name=dict(type='str', aliases=['service', 'unit']),

       #choices里如果没有对应的操作就会报value of state must be one of: reloaded, restarted, started, stopped, got: status
       #报错文件位于ansible/module_utils/basic.py 1810行

state=dict(type='str', choices=['reloaded', 'restarted', 'started', 'stopped', 'status']),

enabled=dict(type='bool'),
masked=dict(type='bool'),
daemon_reload=dict(type='bool', default=False, aliases=['daemon-reload']),
user=dict(type='bool', default=False),
no_block=dict(type='bool', default=False),
),
supports_check_mode=True,
required_one_of=[['state', 'enabled', 'masked', 'daemon_reload']],
) systemctl = module.get_bin_path('systemctl', True)
if module.params['user']:
systemctl = systemctl + " --user"
if module.params['no_block']:
systemctl = systemctl + " --no-block"
unit = module.params['name']
rc = 0
out = err = ''
result = dict(
name=unit,
changed=False,
status=dict(),
) for requires in ('state', 'enabled', 'masked'):
if module.params[requires] is not None and unit is None:
module.fail_json(msg="name is also required when specifying %s" % requires) # Run daemon-reload first, if requested
if module.params['daemon_reload'] and not module.check_mode:
(rc, out, err) = module.run_command("%s daemon-reload" % (systemctl))
if rc != 0:
module.fail_json(msg='failure %d during daemon-reload: %s' % (rc, err)) if unit:
found = False
is_initd = sysv_exists(unit)
is_systemd = False # check service data, cannot error out on rc as it changes across versions, assume not found
#这里执行的命令是类似systemctl show zabbix-agent
(rc, out, err) = module.run_command("%s show '%s'" % (systemctl, unit)) if request_was_ignored(out) or request_was_ignored(err):
# fallback list-unit-files as show does not work on some systems (chroot)
# not used as primary as it skips some services (like those using init.d) and requires .service/etc notation
(rc, out, err) = module.run_command("%s list-unit-files '%s'" % (systemctl, unit))
if rc == 0:
is_systemd = True elif rc == 0:
# load return of systemctl show into dictionary for easy access and return
if out:
#把返回值都扔进result['status']去了
result['status'] = parse_systemctl_show(to_native(out).split('\n')) is_systemd = 'LoadState' in result['status'] and result['status']['LoadState'] != 'not-found' # Check for loading error
if is_systemd and 'LoadError' in result['status']:
module.fail_json(msg="Error loading unit file '%s': %s" % (unit, result['status']['LoadError']))
else:
# Check for systemctl command
module.run_command(systemctl, check_rc=True) # Does service exist?
found = is_systemd or is_initd
if is_initd and not is_systemd:
module.warn('The service (%s) is actually an init script but the system is managed by systemd' % unit) # mask/unmask the service, if requested, can operate on services before they are installed
if module.params['masked'] is not None:
# state is not masked unless systemd affirms otherwise
masked = ('LoadState' in result['status'] and result['status']['LoadState'] == 'masked') if masked != module.params['masked']:
result['changed'] = True
if module.params['masked']:
action = 'mask'
else:
action = 'unmask' if not module.check_mode:
(rc, out, err) = module.run_command("%s %s '%s'" % (systemctl, action, unit))
if rc != 0:
# some versions of system CAN mask/unmask non existing services, we only fail on missing if they don't
fail_if_missing(module, found, unit, msg='host') # if module.params['choices'] == 'status': # Enable/disable service startup at boot if requested
if module.params['enabled'] is not None: if module.params['enabled']:
action = 'enable'
else:
action = 'disable' fail_if_missing(module, found, unit, msg='host') # do we need to enable the service?
enabled = False
(rc, out, err) = module.run_command("%s is-enabled '%s'" % (systemctl, unit)) # check systemctl result or if it is a init script
if rc == 0:
enabled = True
elif rc == 1:
# if not a user service and both init script and unit file exist stdout should have enabled/disabled, otherwise use rc entries
if not module.params['user'] and \
is_initd and \
(not out.strip().endswith('disabled') or sysv_is_enabled(unit)):
enabled = True # default to current state
result['enabled'] = enabled # Change enable/disable if needed
if enabled != module.params['enabled']:
result['changed'] = True
if not module.check_mode:
(rc, out, err) = module.run_command("%s %s '%s'" % (systemctl, action, unit))
if rc != 0:
module.fail_json(msg="Unable to %s service %s: %s" % (action, unit, out + err)) result['enabled'] = not enabled # set service state if requested
if module.params['state'] is not None: fail_if_missing(module, found, unit, msg="host") # default to desired state
result['state'] = module.params['state'] # What is current service state?
if 'ActiveState' in result['status']:
action = None
if module.params['state'] == 'started':
if not is_running_service(result['status']):
action = 'start'
elif module.params['state'] == 'stopped':
if is_running_service(result['status']):
action = 'stop'

# patch status
# 如果是status,则判断返回的result['status'],并标记result['state']
          # action其实是个flag,影响的是接下来的操作.完全可以不对action赋值,就不会影响下面的操作
          # 仿照service.py这个patch,result['status']被我删了

elif module.params['state'] == 'status':
if is_running_service(result['status']):
result['state'] = True
else:
result['state'] = False
del result['status']

else:
if not is_running_service(result['status']):
action = 'start'
else:
action = module.params['state'][:-2] # remove 'ed' from restarted/reloaded
result['state'] = 'started' if action:
result['changed'] = True
if not module.check_mode:
(rc, out, err) = module.run_command("%s %s '%s'" % (systemctl, action, unit))
if rc != 0:
module.fail_json(msg="Unable to %s service %s: %s" % (action, unit, err))
else:
# this should not happen?
module.fail_json(msg="Service is in unknown state", status=result['status']) module.exit_json(**result) if __name__ == '__main__':
main()

染红加粗的几行,就是对status的支持和我写的几行简单说明.

我在git对源码进行了fork,见https://github.com/RedheatWei/ansible-modules-core

对ansible不支持service模块的status命令进行修正的更多相关文章

  1. Ansible简介及常用模块

    一.基础介绍 1.简介 ansible是新出现的自动化运维工具,基于Python开发,集合了众多运维工具(puppet.cfengine.chef.func.fabric)的优点,实现了批量系统配置. ...

  2. ANSIBLE安装和常用模块模块使用详细教程

    目录 ANSIBLE安装和各种模块应用功能 安装配置ANSIBLE ANSIBLE使用 ansible-galaxy工具 ansible-pull工具 ansible-playbook ansible ...

  3. ansible使用,常用模块

    使用ansible管理其他主机有两种方式: 1.命令行执行ansible ad-hoc命令 2.把要做的动作行为写入一个文件[playbook脚本],ansible读取脚本自动完成相应的任务. Ans ...

  4. Ansible指令和常用模块使用

    这里文章记录一下ansible的指令选项和常用的模块使用 ansible指令选项 -m:要执行的模块,默认为command -a:模块的参数 -u:ssh连接的用户名,默认用root,ansible. ...

  5. Ansible Ad-Hoc与常用模块

    ansible 执行结果信息–各颜色说明:ansible Ad-Hoc 说明:ansible 如何查看帮助文档与常用模块详解 主机规划 添加用户账号 说明: 1. 运维人员使用的登录账号: 2. 所有 ...

  6. ansible中文手册-基础模块使用

    此篇文章主要是翻译ansible官网文档而来,在里面讲述了如何使用ansible的基础模块,总体感觉比较晦涩,但是后面会写出自己相关实践的文档,从而更加通俗易懂,官网的东西拿来当手册偶尔翻翻也是很不错 ...

  7. mage Ansible学习1 常用模块

    一.Ansible特点 二.Ansible架构 1.core modules实现常用模块 2.Custom modules实现自定义模块 3.Connection Plugins 连接插件,可通过SS ...

  8. Ansible安装及常用模块

    配置文件:/etc/ansible/ansible.cfg 主机列表:/etc/ansible/hosts  安装anslibe  wget -O /etc/yum.repos.d/epel.repo ...

  9. centos7也支持service命令启动服务吗,对于centos7 中的systemctl和旧的service命令的区别和联系

    一.centos7也支持service命令启动服务吗 CentOS 7.0中一个最主要的改变,就是切换到了systemd.它用于替代红帽企业版Linux前任版本中的SysV和Upstart,对系统和服 ...

随机推荐

  1. 深入Spring Security魔幻山谷-获取认证机制核心原理讲解(新版)

    文/朱季谦 本文基于Springboot+Vue+Spring Security框架而写的原创学习笔记,demo代码参考<Spring Boot+Spring Cloud+Vue+Element ...

  2. 【PHP】用了这么久的Laravel框架,你分析过核心架构了没

    Laravel最初的设计是为了面向MVC架构的,它可以满足如事件处理.用户身份验证等各种需求.另外它还有一个由管理数据库强力支持,用于管理模块化和可扩展性代码的软件包管理器. Laravel以其简洁. ...

  3. 算法很美,听我讲完这些Java经典算法包你爱上她

    大家好,我是小羽. 对于编程来说的话,只有掌握了算法才是了解了编程的灵魂,算法对于新手来说的话,属实有点难度,但是以后想有更好的发展,得到更好的进阶的话,对算法进行系统的学习是重中之重的. 对于 Ja ...

  4. 体渲染——Volume

    基本概念   体渲染(Volume),是绘制类似烟.雾.云的效果.这种渲染和之前的表面渲染不同,光线可以在物体内部进行散射. 体渲染的主要特点    1. 可以在物体内部散射.    2. 从进入vo ...

  5. Python:函数解释(面向过程)

    1. 函数概述 在编程的语境下,函数 (function) 是指一个有命名的.执行某个计算的语句序列 (sequence of statements) .函数可以针对某类问题建立了通用解决步骤(算法) ...

  6. 4. Linux-startx命令

    Linux系统startx命令的功能和使用方法 Linux系统命令startx的功能很简单,就是启动X Window的服务这一项,没有其他的了.其实startx命令启动的是xinit,然后再由xini ...

  7. 网络编程Netty入门:ByteBuf分析

    目录 Netty中的ByteBuf优势 NIO使用的ByteBuffer有哪些缺点 ByteBuf的优势和做了哪些增强 ByteBuf操作示例 ByteBuf操作 简单的Demo示例 堆内和堆外内存 ...

  8. Boxes in a Line UVA - 12657

      You have n boxes in a line on the table numbered 1...n from left to right. Your task is to simulat ...

  9. Django 入门范例

    1. Django 介绍 2. Django 环境搭建 3. 模型(Model) 4. 站点管理 5. 视图(View) 6. 模板(Template) 1. Django 介绍 MVC 模型 大部分 ...

  10. Spring Cloud & Alibaba 实战 | 第十二篇: 微服务整合Sentinel的流控、熔断降级,赋能拥有降级功能的Feign新技能熔断,实现熔断降级双剑合璧(JMeter模拟测试)

    目录 一. Sentinel概念 1. 什么是Sentinel? 2. Sentinel功能特性 3. Sentinel VS Hystrix 二. Docker部署Sentinel Dashboar ...