test_creating_multiple_universe_without_offset fails spuriously

[orbeckst]

Expected behavior

Running our test suite should not fail randomly.

Actual behavior

Tests occasionally fail due to test_creating_multiple_universe_without_offset timing out (see below).

=================================== FAILURES ===================================
________________ test_creating_multiple_universe_without_offset ________________
[gw3] linux -- Python 3.12.2 /home/runner/micromamba/envs/mda/bin/python3.12

temp_xtc = PosixPath('/tmp/pytest-of-runner/pytest-0/popen-gw3/test_creating_multiple_univers0/testing.xtc')
ncopies = 3

    def test_creating_multiple_universe_without_offset(temp_xtc, ncopies=3):
        #  test if they can be created without generating
        #  the offset simultaneously.
        #  The tested XTC file is way too short to induce a race scenario but the
        #  test is included as documentation for the scenario that used to create
        #  a problem (see PR #3375 and issues #3230, #1988)
    
        args = (GRO, str(temp_xtc))
>       with multiprocessing.Pool(2) as p:

/home/runner/work/mdanalysis/mdanalysis/testsuite/MDAnalysisTests/parallelism/test_multiprocessing.py:159: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/pool.py:739: in __exit__
    self.terminate()
/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/pool.py:657: in terminate
    self._terminate()
/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/util.py:227: in __call__
    res = self._callback(*self._args, **self._kwargs)
/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/pool.py:732: in _terminate_pool
    p.join()
/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/process.py:149: in join
    res = self._popen.wait(timeout)
/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/popen_fork.py:43: in wait
    return self.poll(os.WNOHANG if timeout == 0.0 else 0)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <multiprocessing.popen_fork.Popen object at 0x7f1dd0786420>, flag = 0

    def poll(self, flag=os.WNOHANG):
        if self.returncode is None:
            try:
>               pid, sts = os.waitpid(self.pid, flag)
E               Failed: Timeout >200.0s

/home/runner/micromamba/envs/mda/lib/python3.12/multiprocessing/popen_fork.py:27: Failed
----------------------------- Captured stdout call -----------------------------
+++++++++++++++++++++++++++++++++++ Timeout ++++++++++++++++++++++++++++++++++++
~~~~~~~~~~~~~~~~~~~~~ Stack of Thread-2 (139766194525760) ~~~~~~~~~~~~~~~~~~~~~~
  File "/home/runner/micromamba/envs/mda/lib/python3.12/threading.py", line 1030, in _bootstrap
    self._bootstrap_inner()
  File "/home/runner/micromamba/envs/mda/lib/python3.12/threading.py", line 1073, in _bootstrap_inner
    self.run()
  File "/home/runner/micromamba/envs/mda/lib/python3.12/site-packages/tqdm/_monitor.py", line 60, in run
    self.was_killed.wait(self.sleep_interval)
  File "/home/runner/micromamba/envs/mda/lib/python3.12/threading.py", line 655, in wait
    signaled = self._cond.wait(timeout)
  File "/home/runner/micromamba/envs/mda/lib/python3.12/threading.py", line 359, in wait
    gotit = waiter.acquire(True, timeout)
~~~~~~~~~~~~~~~~~~~~~ Stack of <unknown> (139766989321792) ~~~~~~~~~~~~~~~~~~~~~
  File "/home/runner/micromamba/envs/mda/lib/python3.12/site-packages/execnet/gateway_base.py", line 361, in _perform_spawn
    reply.run()
  File "/home/runner/micromamba/envs/mda/lib/python3.12/site-packages/execnet/gateway_base.py", line 296, in run
    self._result = func(*args, **kwargs)
  File "/home/runner/micromamba/envs/mda/lib/python3.12/site-packages/execnet/gateway_base.py", line 1049, in _thread_receiver
    msg = Message.from_io(io)
  File "/home/runner/micromamba/envs/mda/lib/python3.12/site-packages/execnet/gateway_base.py", line 507, in from_io
    header = io.read(9)  # type 1, channel 4, payload 4
  File "/home/runner/micromamba/envs/mda/lib/python3.12/site-packages/execnet/gateway_base.py", line 474, in read
    data = self._read(numbytes - len(buf))
+++++++++++++++++++++++++++++++++++ Timeout ++++++++++++++++++++++++++++++++++++

Solution

Looking at the test

mdanalysis/testsuite/MDAnalysisTests/parallelism/test_multiprocessing.py

Lines 151 to 165 in 0ff14c7

	def test_creating_multiple_universe_without_offset(temp_xtc, ncopies=3):
	# test if they can be created without generating
	# the offset simultaneously.
	# The tested XTC file is way too short to induce a race scenario but the
	# test is included as documentation for the scenario that used to create
	# a problem (see PR #3375 and issues #3230, #1988)
	args = (GRO, str(temp_xtc))
	with multiprocessing.Pool(2) as p:
	universes = [p.apply_async(mda.Universe, args) for i in range(ncopies)]
	universes = [universe.get() for universe in universes]
	assert_equal(universes[0].trajectory._xdr.offsets,
	universes[1].trajectory._xdr.offsets)

we should really first understand if our underlying code is broken or if the test is inherently flaky.

If the underlying code is broken (check issues #3230, #1988 and PR #3375) then we need to fix the code.

If this is an issue with the test then we could add a much shorter timeout using the pytest timeout plugin (which we are already using).

Current version of MDAnalysis

• Which version are you using? (run python -c "import MDAnalysis as mda; print(mda.__version__)") develop 2.8.0-dev on GH CI
• Which version of Python (python -V)? any tested (I think...)
• Which operating system? Linux (macOS??)

[IAlibay]

@orbeckst duplicate of #4209

Will close in favour of that one.