nova-api源码分析(WSGI server的创建及启动)

源码版本:H版

一、前奏

       nova api本身作为一个WSGI服务器,对外提供HTTP请求服务,对内调用nova的其他模块响应相应的HTTP请求。分为两大部分,一是服务器本身的启动与运行,一是加载的app,这个用来处理请求。

   目录结构如下:

      

      首先,nova api是作为一个WSGI服务,肯定要查看它的启动过程,查看启动脚本/etc/init.d/openstack-nova-api(使用service命令实际在调用/etc/init.d文件夹下的脚本文件)。在/etc/init.d/openstack-nova-api文件中,start选项代表调用脚本/usr/bin/nova-api,如下:

 /usr/bin/nova-api

import sys
from nova.cmd.api import main
if __name__ == "__main__":
  sys.exit(main())

接着,分析如下:

/nova/cmd/api.py (文件位置为/usr/lib/python2.6/site-packages,此处及下面均省略)

def main():
    config.parse_args(sys.argv)
    logging.setup("nova")
    utils.monkey_patch()

    launcher = service.process_launcher()
    for api in CONF.enabled_apis:
        should_use_ssl = api in CONF.enabled_ssl_apis
        if api == 'ec2':
            """为了与ec2兼容"""
            server = service.WSGIService(api, use_ssl=should_use_ssl,
                                         max_url_len=16384) 
        else:
            server = service.WSGIService(api, use_ssl=should_use_ssl) 【1
        launcher.launch_service(server, workers=server.workers or 1)【2】
    """当前进程处于等待状态"""
  launcher.wait()

 二、分析【1】处:主要是WSGIService对象的创建

/nova/service.py

WSGIService类:
def __init__(self, name, loader=None, use_ssl=False, max_url_len=None):
    self.name = name
    self.manager = self._get_manager()
    self.loader = loader or wsgi.Loader()
    self.app = self.loader.load_app(name)
    self.host = getattr(CONF, '%s_listen' % name, "0.0.0.0")
    self.port = getattr(CONF, '%s_listen_port' % name, 0)
    self.workers = getattr(CONF, '%s_workers' % name, None)
    self.use_ssl = use_ssl
    """封装了wsgi.py中的server"""
    self.server = wsgi.Server(name,
                              self.app,
                              host=self.host,
                              port=self.port,
                              use_ssl=self.use_ssl,
                              max_url_len=max_url_len)
    # Pull back actual port used
    self.port = self.server.port
    self.backdoor_port = None

/nova/wsgi.py

Loader类:
def __init__(self, config_path=None):
  """试图寻找api-paste.ini配置文件"""
  self.config_path = None

  config_path = config_path or CONF.api_paste_config
  if not os.path.isabs(config_path):
    self.config_path = CONF.find_file(config_path)
  elif os.path.exists(config_path):
    self.config_path = config_path

  if not self.config_path:
    raise exception.ConfigNotFound(path=config_path)

def load_app(self, name):
  try:
    LOG.debug("Loading app %(name)s from %(path)s",
                {'name': name, 'path': self.config_path})
    return deploy.loadapp("config:%s" % self.config_path, name=name)
  except LookupError as err:
    LOG.error(err)
    raise exception.PasteAppNotFound(name=name, path=self.config_path)    

(此处关于app的具体加载过程见后文。)

三、分析【2】处:主要是WSGIService的启动

/nova/openstack/common/service.py

ProcessLauncher类:
def launch_service(self, service, workers=1):
  wrap = ServiceWrapper(service, workers)

  LOG.info(_('Starting %d workers'), wrap.workers)
  """循环启动workers数目的子进程"""
  while self.running and len(wrap.children) < wrap.workers:
    self._start_child(wrap)

def _start_child(self, wrap):
  ...
  """创建子进程"""
  pid = os.fork()
  if pid == 0:
            
    status = 0
    try:
      """子进程进行处理"""
      self._child_process(wrap.service)
    except SignalExit as exc:
      signame = {signal.SIGTERM: 'SIGTERM',
                    signal.SIGINT: 'SIGINT'}[exc.signo]
      LOG.info(_('Caught %s, exiting'), signame)
      status = exc.code
    except SystemExit as exc:
      status = exc.code
    except BaseException:
      LOG.exception(_('Unhandled exception'))
      status = 2
    finally:
      """出现异常停掉服务"""
      wrap.service.stop()
    """子进程退出"""
    os._exit(status)

  LOG.info(_('Started child %d'), pid)
  wrap.children.add(pid)
  self.children[pid] = wrap
  return pid



def _child_process(self, service):
  ...
  launcher = Launcher()
  launcher.run_service(service)

/nova/openstack/common/service.py

Launcher类:
def __init__(self):
    self._services = threadgroup.ThreadGroup()
    self.backdoor_port = eventlet_backdoor.initialize_if_enabled() 


@staticmethod
def run_service(service):
    service.start()
    service.wait()

由于这里的service是WSGIService对象,如下:

/nova/service.py

WSGIService类:
def start(self):
if self.manager: self.manager.init_host() self.manager.pre_start_hook() if self.backdoor_port is not None: self.manager.backdoor_port = self.backdoor_port self.server.start() if self.manager: self.manager.post_start_hook() def wait(self): self.server.wait()

/nova/wsgi.py

Server类:
def start(self):
    ...
    wsgi_kwargs = {
        'func': eventlet.wsgi.server,
        'sock': self._socket,
        'site': self.app,
        'protocol': self._protocol,
        'custom_pool': self._pool,
        'log': self._wsgi_logger,
        'log_format': CONF.wsgi_log_format
        }

    if self._max_url_len:
        wsgi_kwargs['url_length_limit'] = self._max_url_len
    """创建一个green thread,运行func函数,在其中传入后面的参数。此处为调用eventlet.wsgi.server函数,传入sock=self._socket等参数到server函数中。其中eventlet.wsgi.server函数负责启动一个wsgi服务器程序接受客户端发来的请求,
可以参考:http://eventlet.net/doc/modules/wsgi.html?highlight=wsgi.server#eventlet.wsgi.server
""" self._server = eventlet.spawn(**wsgi_kwargs) def wait(self): try: """green thread阻塞等待""" self._server.wait() except greenlet.GreenletExit: LOG.info(_("WSGI server has stopped."))

总结:

      我们可以通过pstree命令查看系统中进程的子进程数,发现nova-api进程中有一个主进程,其子进程数目为一个ec2,一个metadata,有workers数目的osapi_compute(其中workers可以通过/etc/nova/nova.conf中osapi_compute_workers选项设置)。然后通过ps –Lf查看每个进程的线程数目,发现其均为单线程。由此可以看出,整个server的创建和启动过程就是,主进程产生若干的子进程,子进程使用green thread启动wsgi server并等待服务。

参考文档:

http://www.choudan.net/2013/12/09/OpenStack-WSGI-APP%E5%AD%A6%E4%B9%A0.html

http://developer.openstack.org/api-ref-compute-v2.html(nova的API文档)

posted on 2014-10-13 20:16  小杰鱼  阅读(3590)  评论(0编辑  收藏  举报

导航