使用das提交任务时出现Pickle错误

Question

使用一个简单的dask方法执行一个简单的dask方法失败。在

有人能给我指出正确的方向吗。在

下面是我正在运行的代码：

from dask.distributed import Client, as_completed
import time

class DaskConnect:

def __init__(self):
    print("Initialized:",self.__class__.__name__)
    self.scheduler_host="192.168.0.4"
    self.scheduler_port="8786"

def connect(self):
    self.client = Client(self.scheduler_host+":"+self.scheduler_port)
    # self.client = Client()
    return self.client

def disconnect(self):
    self.client.close()

class TestDask:
def __init__(self):
    print("Initialized:",self.__class__.__name__)
    self.dask_client=DaskConnect().connect()

def do_task(self,msg):
    time.sleep(30)
    return msg

def run(self):
    tasks=[1]
    # tasks = [1, 2, 3, 4, 5]
    futures=[]
    for task in tasks:
        print("Submitting:",task)
        future = self.dask_client.submit(self.do_task, "Task:"+str(task))
        futures.append(future)

    for future in as_completed(futures):
        result = future.result()
        print("Result",result)

TestDask().run()

错误：

distributed.protocol.pickle - INFO - Failed to serialize main.TestDask object at 0x101c408d0>>. Exception: can't pickle select.kqueue objects Traceback (most recent call last):

Answer 1

Dask客户端当前不可序列化。任何包含Dask客户机的对象也不能序列化。通常，序列化包含活动网络连接、锁等的任何内容都是一个挑战。。在

也许你的问题还有别的办法？在