[BUG] Pymars shows ambiguous error messages and incorrect progress bar

[ChengjieLi28]

Describe the bug
When a job failed due to maybe OOM, pymars shows ambiguous error messages and incorrect progress bar.

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

1. Your Python version 3.8.5
2. The version of Mars you use: latest
3. Versions of crucial packages, such as numpy, scipy and pandas: follow pymars
4. Full stack of the error.

Traceback (most recent call last):
  File "/opt/mars/benchmarks/tpch/run_queries.py", line 1063, in <module>
    main()
  File "/opt/mars/benchmarks/tpch/run_queries.py", line 1056, in main
    run_queries(folder, use_arrow_dtype=use_arrow_dtype)
  File "/opt/mars/benchmarks/tpch/run_queries.py", line 986, in run_queries
    mars.execute([lineitem, orders, customer, nation, region, supplier, part, partsupp])
  File "/opt/mars/mars/deploy/oscar/session.py", line 1890, in execute
    return session.execute(
  File "/opt/mars/mars/deploy/oscar/session.py", line 1684, in execute
    execution_info: ExecutionInfo = fut.result(
  File "/opt/conda/lib/python3.8/concurrent/futures/_base.py", line 439, in result
    return self.__get_result()
  File "/opt/conda/lib/python3.8/concurrent/futures/_base.py", line 388, in __get_result
    raise self._exception
  File "/opt/mars/mars/deploy/oscar/session.py", line 1870, in _execute
    await execution_info
  File "/opt/mars/mars/deploy/oscar/session.py", line 105, in wait
    return await self._aio_task
  File "/opt/mars/mars/deploy/oscar/session.py", line 953, in _run_in_background
    raise task_result.error.with_traceback(task_result.traceback)
  File "/opt/mars/mars/services/task/supervisor/processor.py", line 369, in run
    await self._process_stage_chunk_graph(*stage_args)
  File "/opt/mars/mars/services/task/supervisor/processor.py", line 247, in _process_stage_chunk_graph
    chunk_to_result = await self._executor.execute_subtask_graph(
  File "/opt/mars/mars/services/task/execution/mars/executor.py", line 196, in execute_subtask_graph
    return await stage_processor.run()
  File "/opt/mars/mars/services/task/execution/mars/stage.py", line 240, in run
    raise self.result.error.with_traceback(self.result.traceback)
  File "/opt/mars/mars/services/scheduling/worker/execution.py", line 392, in internal_run_subtask
    subtask_info.result = await self._retry_run_subtask(
  File "/opt/mars/mars/services/scheduling/worker/execution.py", line 497, in _retry_run_subtask
    return await _retry_run(subtask, subtask_info, _run_subtask_once)
  File "/opt/mars/mars/services/scheduling/worker/execution.py", line 91, in _retry_run
    raise ex
  File "/opt/mars/mars/services/scheduling/worker/execution.py", line 69, in _retry_run
    return await target_async_func(*args)
  File "/opt/mars/mars/services/scheduling/worker/execution.py", line 475, in _run_subtask_once
    raise ex
  File "/opt/mars/mars/services/scheduling/worker/execution.py", line 439, in _run_subtask_once
    return await asyncio.shield(aiotask)
  File "/opt/mars/mars/services/subtask/api.py", line 68, in run_subtask_in_slot
    return await ref.run_subtask.options(profiling_context=profiling_context).send(
  File "/opt/mars/mars/oscar/backends/context.py", line 195, in send
    result = await self._wait(future, actor_ref.address, message)
  File "/opt/mars/mars/oscar/backends/context.py", line 89, in _wait
    return await future
  File "/opt/mars/mars/oscar/backends/context.py", line 80, in _wait
    await asyncio.shield(future)
  File "/opt/mars/mars/oscar/backends/core.py", line 68, in _listen
    raise ServerClosed(
mars.oscar.errors.ServerClosed: Remote server unixsocket:///917504 closed

5. Minimized code to reproduce the error.

 python mars/benchmarks/tpch/run_queries.py --folder /opt/data-1G/ --query 1 --endpoint http://10.0.0.4:8001

Should run this tpch query in a env with less memory to trigger this error.

My Azure VM is 2 core and 8G memory. When running query 1 with tpch data 1G, this error happens.
The mars UI shows that 100% progress but failed.