Server

rpyc plug-in server (threaded or forking)

class rpyc.utils.server.Server(service, hostname=None, ipv6=False, port=0, backlog=128, reuse_addr=True, authenticator=None, registrar=None, auto_register=None, protocol_config=None, logger=None, listener_timeout=0.5, socket_path=None)

Base server implementation

Parameters:

• service – the Service to expose

• hostname – the host to bind to. By default, the ‘wildcard address’ is used to listen on all interfaces. if not properly secured, the server can receive traffic from unintended or even malicious sources.

• ipv6 – whether to create an IPv6 or IPv4 socket. The default is IPv4

• port – the TCP port to bind to

• backlog – the socket’s backlog (passed to listen())

• reuse_addr – whether or not to create the socket with the SO_REUSEADDR option set.

• authenticator – the Authenticators to use. If None, no authentication is performed.

• registrar – the RegistryClient to use. If None, a default UDPRegistryClient will be used

• auto_register – whether or not to register using the registrar. By default, the server will attempt to register only if a registrar was explicitly given.

• protocol_config – the configuration dictionary that is passed to the RPyC connection

• logger – the logger to use (of the built-in logging module). If None, a default logger will be created.

• listener_timeout – the timeout of the listener socket; set to None to disable (e.g. on embedded platforms with limited battery)

close()

Closes (terminates) the server and all of its clients. If applicable, also unregisters from the registry server

fileno()

returns the listener socket’s file descriptor

accept()

accepts an incoming socket connection (blocking)

start()

Starts the server (blocking). Use close() to stop

class rpyc.utils.server.OneShotServer(service, hostname=None, ipv6=False, port=0, backlog=128, reuse_addr=True, authenticator=None, registrar=None, auto_register=None, protocol_config=None, logger=None, listener_timeout=0.5, socket_path=None)

A server that handles a single connection (blockingly), and terminates after that

Parameters: see Server

class rpyc.utils.server.ThreadedServer(service, hostname=None, ipv6=False, port=0, backlog=128, reuse_addr=True, authenticator=None, registrar=None, auto_register=None, protocol_config=None, logger=None, listener_timeout=0.5, socket_path=None)

A server that spawns a thread for each connection. Works on any platform that supports threads.

Parameters: see Server

class rpyc.utils.server.ThreadPoolServer(*args, **kwargs)

This server is threaded like the ThreadedServer but reuses threads so that recreation is not necessary for each request. The pool of threads has a fixed size that can be set with the ‘nbThreads’ argument. The default size is 20. The server dispatches request to threads by batch, that is a given thread may process up to request_batch_size requests from the same connection in one go, before it goes to the next connection with pending requests. By default, self.request_batch_size is set to 10 and it can be overwritten in the constructor arguments.

Contributed by @sponce

Parameters: see Server

close()

closes a ThreadPoolServer. In particular, joins the thread pool.

class rpyc.utils.server.ForkingServer(*args, **kwargs)

A server that forks a child process for each connection. Available on POSIX compatible systems only.

Parameters: see Server

close()

Closes (terminates) the server and all of its clients. If applicable, also unregisters from the registry server

class rpyc.utils.server.GeventServer(service, hostname=None, ipv6=False, port=0, backlog=128, reuse_addr=True, authenticator=None, registrar=None, auto_register=None, protocol_config=None, logger=None, listener_timeout=0.5, socket_path=None)

gevent based Server. Requires using gevent.monkey.patch_all().