[BUG] Ray DAG mode access Mars WEB Dashboard error

[chaokunyang]

Describe the bug When access task status for ray DAG mode in mars dashboard, got incorrect task status. Following task is finished, the graph should be green instead of blank:

To Reproduce To help us reproducing this bug, please provide information below:

1. Your Python version: 3.8
2. The version of Mars you use: https://github.com/mars-project/mars/pull/3165
3. Versions of crucial packages, such as numpy, scipy and pandas: pandas 1.4.2, numpy 1.19.5, scipy 1.8.1
4. Full stack of the error.

(RayMainPool pid=54669) 2022-06-27 16:06:15,283 ERROR web.py:2239 -- 500 GET /api/session/SFJWHJesbcMFjANMcqTskZ6R/task/iQvlLWn2zC9z63HSWWbw5maQ/tileable_detail (127.0.0.1) 5.22ms
^C(RayMainPool pid=54669) 2022-06-27 16:06:16,283       ERROR core.py:82 -- TypeError when handling request with TaskWebAPIHandler.get_tileable_details
(RayMainPool pid=54669) Traceback (most recent call last):
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/web/core.py", line 70, in wrapped
(RayMainPool pid=54669)     res = await self._create_or_get_url_future(
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/task/api/web.py", line 132, in get_tileable_details
(RayMainPool pid=54669)     res = await oscar_api.get_tileable_details(task_id)
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/task/api/oscar.py", line 77, in get_tileable_details
(RayMainPool pid=54669)     return await self._task_manager_ref.get_tileable_details(task_id)
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 263, in __pyx_actor_method_wrapper
(RayMainPool pid=54669)     async with lock:
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 266, in mars.oscar.core.__pyx_actor_method_wrapper
(RayMainPool pid=54669)     result = await result
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/task/supervisor/manager.py", line 206, in get_tileable_details
(RayMainPool pid=54669)     return await processor_ref.get_tileable_details()
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/oscar/backends/context.py", line 196, in send
(RayMainPool pid=54669)     return self._process_result_message(result)
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/oscar/backends/context.py", line 76, in _process_result_message
(RayMainPool pid=54669)     raise message.as_instanceof_cause()
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/oscar/backends/pool.py", line 586, in send
(RayMainPool pid=54669)     result = await self._run_coro(message.message_id, coro)
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/oscar/backends/pool.py", line 343, in _run_coro
(RayMainPool pid=54669)     return await coro
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/oscar/api.py", line 120, in __on_receive__
(RayMainPool pid=54669)     return await super().__on_receive__(message)
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 523, in __on_receive__
(RayMainPool pid=54669)     raise ex
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 516, in mars.oscar.core._BaseActor.__on_receive__
(RayMainPool pid=54669)     return await self._handle_actor_result(result)
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 401, in _handle_actor_result
(RayMainPool pid=54669)     task_result = await coros[0]
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 444, in mars.oscar.core._BaseActor._run_actor_async_generator
(RayMainPool pid=54669)     async with self._lock:
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 445, in mars.oscar.core._BaseActor._run_actor_async_generator
(RayMainPool pid=54669)     with debug_async_timeout('actor_lock_timeout',
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 450, in mars.oscar.core._BaseActor._run_actor_async_generator
(RayMainPool pid=54669)     res = await gen.athrow(*res)
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/task/supervisor/task.py", line 159, in get_tileable_details
(RayMainPool pid=54669)     tileable_to_details = yield asyncio.to_thread(self._get_tileable_infos)
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 455, in mars.oscar.core._BaseActor._run_actor_async_generator
(RayMainPool pid=54669)     res = await self._handle_actor_result(res)
(RayMainPool pid=54669)   File "mars/oscar/core.pyx", line 375, in _handle_actor_result
(RayMainPool pid=54669)     result = await result
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/lib/aio/_threads.py", line 36, in to_thread
(RayMainPool pid=54669)     return await loop.run_in_executor(None, func_call)
(RayMainPool pid=54669)   File "/Users/chaokunyang/anaconda3/envs/mars3.8/lib/python3.8/concurrent/futures/thread.py", line 57, in run
(RayMainPool pid=54669)     result = self.fn(*self.args, **self.kwargs)
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/task/supervisor/task.py", line 100, in _get_tileable_infos
(RayMainPool pid=54669)     subtask_id_to_results = self._get_all_subtask_results()
(RayMainPool pid=54669)   File "/Users/chaokunyang/Desktop/chaokun/python/mars/mars/services/task/supervisor/task.py", line 64, in _get_all_subtask_results
(RayMainPool pid=54669)     for stage in processor.stage_processors:
(RayMainPool pid=54669) TypeError: [address=ray://ray-cluster-1656317089/0/0, pid=54669] 'NoneType' object is not iterable
(RayMainPool pid=54669) 2022-06-27 16:06:16,289 ERROR web.py:2239 -- 500 GET /api/session/SFJWHJesbcMFjANMcqTskZ6R/task/iQvlLWn2zC9z63HSWWbw5maQ/tileable_detail (127.0.0.1) 8.59ms

5. Minimized code to reproduce the error. pytest -v -s mars/deploy/oscar/tests/test_ray_dag_oscar.py::test_iterative_tiling:

@require_ray
@pytest.mark.asyncio
async def test_iterative_tiling(ray_start_regular_shared2, create_cluster):
    await test_local.test_iterative_tiling(create_cluster)
    time.sleep(100000)

Expected behavior A clear and concise description of what you expected to happen.

Additional context Add any other context about the problem here.