当前位置: 首页>>代码示例>>Python>>正文


Python SyncManager.get_nodes方法代码示例

本文整理汇总了Python中multiprocessing.managers.SyncManager.get_nodes方法的典型用法代码示例。如果您正苦于以下问题:Python SyncManager.get_nodes方法的具体用法?Python SyncManager.get_nodes怎么用?Python SyncManager.get_nodes使用的例子?那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在multiprocessing.managers.SyncManager的用法示例。


在下文中一共展示了SyncManager.get_nodes方法的1个代码示例,这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞,您的评价将有助于系统推荐出更棒的Python代码示例。

示例1: Node

# 需要导入模块: from multiprocessing.managers import SyncManager [as 别名]
# 或者: from multiprocessing.managers.SyncManager import get_nodes [as 别名]
class Node(Daemon):
   """
   Node is started up on the remote instance via the bootstrapping process for that instance.
   The node is responsible for tracking active streams and managing the workers that process
   the jobs from thosee streams.  If a stream goes idle (ie, there are no more jobs in the streams
   queue and all workers have died) then node will stop tracking the stream.  If jobs re-appear
   on the stream Node will spawn new workers to process those jobs.  If a new stream appears 
   Node will spawn new workers to processs the jobs on that stream.  Each worker is an independent
   concurrent process that inherits the stream to process from the Node.
   """

   def __init__(self, queue, qauthkey, mpps= 5, dfs= None, dauthkey= None, logdir= curdir, piddir= curdir, **properties):
      """Initialize the Node's I/O stream and connect to the Queue and/or DFS."""     

      self.id= getipaddress()
      self.queue= queue
      self.qauthkey= qauthkey
      self.mpps= mpps
      self.dfs= dfs
      self.dauthkey= dauthkey
      self.properties= properties
      self.shutdown= Value('i', 0)

      self.workers= {}
      self.alive= True
      self.start_time= datetime.utcnow()

      self.connect()
  
      super(Node, self).__init__(
         pidfile= path.join(piddir, self.__class__.__name__ + ".pid"),
         stdout= path.join(logdir, self.__class__.__name__ + ".out"),
         stderr= path.join(logdir, self.__class__.__name__ + ".err"),
         stdin= path.join(logdir, self.__class__.__name__ + ".in")
      )

   def connect(self):
      """Connects to the Queue and/or DFS on the host/port for whic hthe Node was intialized for."""

      self.qconnect()
      if None not in self.dfs:
         self.dconnect()

   def qconnect(self):
      """
      Attempts to connect to the Queue on the host/port for which the Node was initialized for.
      If no connection can be made, Node will keep attempting to connect until a connection
      can be established.  One connection is established the remove methods requested will be
      registered.
      """

      # remove connection from cache:
      # BaseProxy class has thread local storage which caches the connection
      # which is reused for future connections causing "borken pipe" errors on 
      # creating new manager.  
      if self.queue in BaseProxy._address_to_local:
         if hasattr(BaseProxy._address_to_local[self.queue][0], 'connection'):
            del BaseProxy._address_to_local[self.queue][0].connection

      # register handlers
      SyncManager.register("get_streams")
      SyncManager.register("get_queue")
      SyncManager.register("get_store")
      SyncManager.register("get_properties")

      print "connecting to queue", self.queue
      while self.alive:

         try:
            self.impq= SyncManager(address= self.queue, authkey= self.qauthkey)
            self.impq.connect() 
            print "connected to queue", self.queue
            break
         except (EOFError, IOError, SocketError) as e:
            print "could not connect ...trying again", str(e)
            sleep(1)

   def dconnect(self):
      """
      Attempts to connect to the DFS on the host/port for which the Node was initialized for.
      If no connection can be made, Node will keep attempting to connect until a connection
      can be established. Once a connection can be established the remove methods requested
      will be registered.
      """

      # remove connection from cache:
      # BaseProxy class has thread local storage which caches the connection
      # which is reused for future connections causing "borken pipe" errors on
      # creating new manager.
      if self.dfs in BaseProxy._address_to_local:
         if hasattr(BaseProxy._address_to_local[self.dfs][0], 'connection'):
            del BaseProxy._address_to_local[self.dfs][0].connection

      # register handlers
      SyncManager.register("get_nodes")

      print "connecting to dfs", self.dfs
      while self.alive:

         try:
#.........这里部分代码省略.........
开发者ID:richardjmarini,项目名称:Impetus,代码行数:103,代码来源:impetus.py


注:本文中的multiprocessing.managers.SyncManager.get_nodes方法示例由纯净天空整理自Github/MSDocs等开源代码及文档管理平台,相关代码片段筛选自各路编程大神贡献的开源项目,源码版权归原作者所有,传播和使用请参考对应项目的License;未经允许,请勿转载。