openstack Rocky系列之keystone:(二)keystone中API注册
主要说一下initialize_application中的application_factory
def loadapp():
app = application.application_factory(name)
return app
源码如下
@fail_gracefully
def application_factory(name='public'):
if name not in ('admin', 'public'):
raise RuntimeError('Application name (for base_url lookup) must be '
'either `admin` or `public`.') # NOTE(morgan): The Flask App actually dispatches nothing until we migrate
# some routers to Flask-Blueprints, it is simply a placeholder.
app = flask.Flask(name)
app.after_request(_add_vary_x_auth_token_header) # NOTE(morgan): Configure the Flask Environment for our needs.
app.config.update(
# We want to bubble up Flask Exceptions (for now)
PROPAGATE_EXCEPTIONS=True) # TODO(morgan): Convert Subsystems over to Flask-Native, for now, we simply
# dispatch to another "application" [e.g "keystone"]
# NOTE(morgan): as each router is converted to flask-native blueprint,
# remove from this list. WARNING ORDER MATTERS; ordered dict used to
# ensure sane ordering of the routers in the legacy-dispatch model.
dispatch_map = collections.OrderedDict() # Load in Healthcheck and map it to /healthcheck
hc_app = healthcheck.Healthcheck.app_factory(
{}, oslo_config_project='keystone')
dispatch_map['/healthcheck'] = hc_app # More legacy code to instantiate all the magic for the dispatchers.
# The move to blueprints (FLASK) will allow this to be eliminated.
_routers = []
sub_routers = []
mapper = routes.Mapper()
for api_routers in ALL_API_ROUTERS:
moved_found = [pfx for
pfx in getattr(api_routers, '_path_prefixes', [])
if pfx in _MOVED_API_PREFIXES]
if moved_found:
raise RuntimeError('An API Router is trying to register path '
'prefix(s) `%(pfx)s` that is handled by the '
'native Flask app. Keystone cannot '
'start.' %
{'pfx': ', '.join([p for p in moved_found])}) routers_instance = api_routers.Routers()
_routers.append(routers_instance)
routers_instance.append_v3_routers(mapper, sub_routers)
# TODO(morgan): Remove "API version registration". For now this is kept
# for ease of conversion (minimal changes)
keystone.api.discovery.register_version('v3') # NOTE(morgan): We add in all the keystone.api blueprints here, this
# replaces (as they are implemented) the legacy dispatcher work.
for api in keystone.api.__apis__:
for api_bp in api.APIs:
api_bp.instantiate_and_register_to_app(app) # Build and construct the dispatching for the Legacy dispatching model
sub_routers.append(_ComposibleRouterStub(_routers))
legacy_dispatcher = keystone_wsgi.ComposingRouter(mapper, sub_routers) for pfx in itertools.chain(*[rtr.Routers._path_prefixes for
rtr in ALL_API_ROUTERS]):
import pdb;pdb.set_trace()
dispatch_map['/v3/%s' % pfx] = legacy_dispatcher
app.wsgi_app = KeystoneDispatcherMiddleware(
app.wsgi_app,
dispatch_map)
return app
fail_gracefully是个装饰器,主要是为了防止函数报错,抓一下exception,并记录报错信息,其中又使用了一个functools.wraps的装饰器,防止fail_gracefully对被装饰的函数造成影响
主要看27行dispatch_map = collections.OrderedDict(),这里创建了一个dispatch_map的变量,这个变量用于记录所有的路由分发 33行创建了一个router.Mapper()的对象mapper
for api_routers in ALL_API_ROUTERS:
moved_found = [pfx for
pfx in getattr(api_routers, '_path_prefixes', [])
if pfx in _MOVED_API_PREFIXES]
if moved_found:
raise RuntimeError('An API Router is trying to register path '
'prefix(s) `%(pfx)s` that is handled by the '
'native Flask app. Keystone cannot '
'start.' %
{'pfx': ', '.join([p for p in moved_found])}) routers_instance = api_routers.Routers()
_routers.append(routers_instance)
routers_instance.append_v3_routers(mapper, sub_routers)
ALL_API_ROUTER是一个包含所有处理路由的类的列表,这个循环中主要是过滤掉所有的被移除的API,api_routes.Routers()继承wsgi.RoutersBase,并通过append_v3_routers将自己的路由信息记录到mapper中,其中sub_routers是记录顶层路由下的路由
sub_routers.append(_ComposibleRouterStub(_routers))这行代码将_router整合到sub_routers中
legacy_dispatcher = keystone_wsgi.ComposingRouter(mapper, sub_routers)这行代码最终调用的源代码如下
class Router(wsgi.ComposableRouter):
def __init__(self, controller, collection_key, key,
resource_descriptions=None,
is_entity_implemented=True,
method_template=None):
self.controller = controller
self.key = key
self.collection_key = collection_key
self._resource_descriptions = resource_descriptions
self._is_entity_implemented = is_entity_implemented
self.method_template = method_template or '%s' def add_routes(self, mapper):
collection_path = '/%(collection_key)s' % {
'collection_key': self.collection_key}
entity_path = '/%(collection_key)s/{%(key)s_id}' % {
'collection_key': self.collection_key,
'key': self.key} mapper.connect(
collection_path,
controller=self.controller,
action=self.method_template % 'create_%s' % self.key,
conditions=dict(method=['POST']))
mapper.connect(
collection_path,
controller=self.controller,
action=self.method_template % 'list_%s' % self.collection_key,
conditions=dict(method=['GET', 'HEAD']))
mapper.connect(
entity_path,
controller=self.controller,
action=self.method_template % 'get_%s' % self.key,
conditions=dict(method=['GET', 'HEAD']))
mapper.connect(
entity_path,
controller=self.controller,
action=self.method_template % 'update_%s' % self.key,
conditions=dict(method=['PATCH']))
mapper.connect(
entity_path,
controller=self.controller,
action=self.method_template % 'delete_%s' % self.key,
conditions=dict(method=['DELETE'])) # Add the collection resource and entity resource to the resource
# descriptions. collection_rel = json_home.build_v3_resource_relation(
self.collection_key)
rel_data = {'href': collection_path, }
self._resource_descriptions.append((collection_rel, rel_data))
json_home.JsonHomeResources.append_resource(collection_rel, rel_data) if self._is_entity_implemented:
entity_rel = json_home.build_v3_resource_relation(self.key)
id_str = '%s_id' % self.key
id_param_rel = json_home.build_v3_parameter_relation(id_str)
entity_rel_data = {
'href-template': entity_path,
'href-vars': {
id_str: id_param_rel,
},
}
self._resource_descriptions.append((entity_rel, entity_rel_data))
json_home.JsonHomeResources.append_resource(
entity_rel, entity_rel_data)
通过add_router为route添加mapper
for pfx in itertools.chain(*[rtr.Routers._path_prefixes for
rtr in ALL_API_ROUTERS]):
dispatch_map['/v3/%s' % pfx] = legacy_dispatcher
这里主要添加dispatch_map中的前缀和处理代码的映射关系,涉及到itertools.chain()用法的小技巧

最后一行的KeystoneDispatcherMiddleware类负责API的分发,同样是通过__call__()这个魔法函数
openstack Rocky系列之keystone:(二)keystone中API注册的更多相关文章
- openstack Rocky系列之keystone:(一)keystone的启动
keystone在httpd的入口执行文件为/usr/bin/keystone-wsgi-public 查看文件/usr/bin/keystone-wsgi-public,根据代码,看到主要是这几行代 ...
- OpenStack学习系列之十二:安装ceph并对接OpenStack
Ceph 是一种为优秀的性能.可靠性和可扩展性而设计的统一的.分布式文件系统.Ceph 的统一体现在可以提供文件系统.块存储和对象存储,分布式体现在可以动态扩展.在国内一些公司的云环境中,通常 ...
- 我非要捅穿这 Neutron(三)架构分析与代码实现篇(基于 OpenStack Rocky)
目录 文章目录 目录 Neutron 的软件架构分析与实现 Neutron Server 启动流程 获取 WSGI Application Core API & Extension API C ...
- OpenStack实践系列②认证服务Keystone
OpenStack实践系列②认证服务Keystone 三.实战OpenStack之控制节点3.1 CentOS7的时间同步服务器chrony 下载chrony # yum install -y chr ...
- (原创)OpenStack服务如何使用Keystone (二)---部署和配置Keystone中间件
(一)Keystone端的操作 (二)如何在OpenStack服务上部署Keystone中间件 (三)详细配置keystonemiddleware 部署OpenStack时一般先安装Keystone服 ...
- OpenStack组件系列☞Keystone搭建
一:版本信息 官网:http://docs.openstack.org/newton/install-guide-rdo/keystone.html 二:部署keystone 官网文档:http:// ...
- OpenStack组件系列☞Keystone
Keystone(OpenStack Identity Service)是 OpenStack 框架中负责管理身份验证.服务规则和服务令牌功能的模块.用户访问资源需要验证用户的身份与权限,服务执行操作 ...
- openstack项目【day23】:keystone组件基础
本节内容 一 什么是keystone 二 为何要有keystone 三 keystone的功能 四 keystone概念详解 五 keystone内包含的组件 六 keystone与openstack ...
- openstack项目【day24】:keystone部署及操作
阅读目录 一 前言 二 版本信息 三 部署keystone 四 keystone操作 五 验证 六 创建脚本 七 keystone使用套路总结 一 前言 任何软件的部署都是没有技术含量的,任何就部署讲 ...
随机推荐
- linux 执行:pip3 install -r requirements.txt 报错
错误内容: 解决办法:
- Python学习小技巧之列表项的排序
Python学习小技巧之列表项的排序 本文介绍的是关于Python列表项排序的相关内容,分享出来供大家参考学习,下面来看看详细的介绍: 典型代码1: data_list = [6, 9, 1, ...
- linux之用户,用户组,软件操作
用户 超级管理员用户 root 0 普通用户 (0-65535) 系统用户:启动系统服务和进程的用户,不可以登陆. (1-999centos7)(1-499centos6) 可登陆用户:能登录系统的用 ...
- vue 使用 npm run dev命令后 自动打开浏览器
1.使用vue-cli 老版本构建项目时, 可修改config文件夹下index.js文件 autoOpenBrowser 属性给为 true 即可 使用vue-cli 3.x 版本后,所有的配置项均 ...
- C# Job System概述
C# 作业系统是如何工作 允许用户编写与Unity其余部分良好交互的多线程代码,并使编写正确的代码变得更加容易. 编写多线程代码可以提供高性能的好处,其中包括帧速率的显著提高.以及将Burst编译器与 ...
- base64图片数据类型转numpy的ndarray矩阵类型数据
1.两种方法如下链接 https://www.cnblogs.com/mtcnn/p/9411683.html 2.第一种方法: # coding: utf-8 # python base64 编解码 ...
- 【Python开发】使用pyplot模块绘图
快速绘图 使用pyplot模块绘图¶ matplotlib的pyplot模块提供了和MATLAB类似的绘图API,方便用户快速绘制二维图表.我们先看一个简单的例子: 05-matplotlib/mat ...
- Pandas时间序列和分组聚合
#时间序列import pandas as pd import numpy as np # 生成一段时间范围 ''' 该函数主要用于生成一个固定频率的时间索引,在调用构造方法时,必须指定start.e ...
- Python学习【day04】- Python基础(集合、函数)
集合 #!/usr/bin/env python # -*- coding:utf8 -*- # set集合 只可放不可变的数据类型,本身是可变数据类型,无序 # s = {1,2,3,[1,2,3] ...
- PHP空对象 空数组
PHP定义空对象:$obj = (object)null;或$obj = (object)array(); 定义空数组:$arr = array();或$arr = [];//自 5.4 起可以使用短 ...