ClusterRuntime is a runtime service that manages dynamic configuration and status of the whole cluster. More...

#include <cluster_rt.h>

Public Member Functions
	ClusterRuntime (const std::string &host, int job_id)

	ClusterRuntime (const std::string &host, int job_id, int timeout)

bool	Init ()
	Initialize the runtime instance.

int	RegistProc (const std::string &host_addr, int pid)
	register the process, and get a unique process id More...

std::string	GetProcHost (int proc_id)
	translate the process id to host address More...

bool	WatchSGroup (int gid, int sid, rt_callback fn, void *ctx)
	Server: watch all workers in a server group, will be notified when all workers have left.

bool	JoinSGroup (int gid, int wid, int s_group)
	Worker: join a server group (i.e. More...

bool	LeaveSGroup (int gid, int wid, int s_group)
	Worker: leave a server group (i.e. More...

Detailed Description

ClusterRuntime is a runtime service that manages dynamic configuration and status of the whole cluster.

It mainly provides following services: 1) Provide running status of each server/worker 2) Translate process id to (hostname:port)

Member Function Documentation

std::string singa::ClusterRuntime::GetProcHost ( int proc_id )

translate the process id to host address

bool singa::ClusterRuntime::JoinSGroup	(	int	gid,
		int	wid,
		int	s_group
	)

Worker: join a server group (i.e.

start to read/update these servers)

bool singa::ClusterRuntime::LeaveSGroup	(	int	gid,
		int	wid,
		int	s_group
	)

Worker: leave a server group (i.e.

finish its all work)

int singa::ClusterRuntime::RegistProc	(	const std::string &	host_addr,
		int	pid
	)

register the process, and get a unique process id

The documentation for this class was generated from the following file:

/home/wangwei/program/asf/release-0.1/apache-singa-incubating-0.1.0-RC1/include/utils/cluster_rt.h