zoo.ray package¶
Submodules¶
zoo.ray.process module¶
-
class
zoo.ray.process.ProcessInfo(out, err, errorcode, pgid, tag='default', pids=None, node_ip=None)[source]¶ Bases:
object
zoo.ray.raycontext module¶
-
class
zoo.ray.raycontext.JVMGuard[source]¶ Bases:
objectThe registered pids would be put into the killing list of Spark Executor.
-
class
zoo.ray.raycontext.RayContext(sc, redis_port=None, password='123456', object_store_memory=None, verbose=False, env=None, extra_params=None, num_ray_nodes=None, ray_node_cpu_cores=None)[source]¶ Bases:
object-
address_info¶
-
init(driver_cores=0)[source]¶ Initiate the ray cluster.
Parameters: driver_cores – The number of cores for the raylet on driver for Spark cluster mode. Default is 0 and in this case the local driver wouldn’t have any ray workload.
:return The dictionary of address information about the ray cluster. Information contains node_ip_address, redis_address, object_store_address, raylet_socket_name, webui_url and session_dir.
-
-
class
zoo.ray.raycontext.RayServiceFuncGenerator(python_loc, redis_port, ray_node_cpu_cores, password, object_store_memory, verbose=False, env=None, extra_params=None)[source]¶ Bases:
objectThis should be a pickable class.
-
zoo.ray.raycontext.kill_redundant_log_monitors(redis_address)[source]¶ Killing redundant log_monitor.py processes. If multiple ray nodes are started on the same machine, there will be multiple ray log_monitor.py processes monitoring the same log dir. As a result, the logs will be replicated multiple times and forwarded to driver. See issue https://github.com/ray-project/ray/issues/10392