Bug#1081459: FTBFS: test_util_pool fails

Drew Parsons dparsons at debian.org
Wed Sep 11 19:56:33 BST 2024


Package: python3-mpi4py
Version: 4.0.0-2
Severity: serious
Tags: ftbfs
Justification: FTBFS
Control: forwarded -1 https://github.com/mpi4py/mpi4py/issues/545

mpi4py 4 was building fine in experimental but shows tests error in
unstable:

testTestAll (test_util_pkl5.TestPKL5World.testTestAll) ... ok
testWaitAll (test_util_pkl5.TestPKL5World.testWaitAll) ... ok
test_apply (test_util_pool.TestProcessPool.test_apply) ... [sbuild:00243] 1 more process has sent help message help-mca-bml-r2.txt / unreachable proc
[sbuild:00483] [[13078,42],0] ORTE_ERROR_LOG: Unreachable in file ../../../ompi/dpm/dpm.c at line 493
[sbuild:00242] [[13078,1],0] ORTE_ERROR_LOG: Unreachable in file ../../../ompi/dpm/dpm.c at line 493
Exception in thread Thread-6 (_manager_spawn):
Traceback (most recent call last):
  File "/usr/lib/python3.12/threading.py", line 1075, in _bootstrap_inner
    self.run()
  File "/usr/lib/python3.12/threading.py", line 1012, in run
    self._target(*self._args, **self._kwargs)
  File "/<<PKGBUILDDIR>>/.pybuild/cpython3_3.12/build/mpi4py/futures/_core.py", line 350, in _manager_spawn
    comm = serialized(client_spawn)(pyexe, pyargs, nprocs, info)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/<<PKGBUILDDIR>>/.pybuild/cpython3_3.12/build/mpi4py/futures/_core.py", line 1058, in client_spawn
    comm = MPI.COMM_SELF.Spawn(python_exe, args, max_workers, info)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "src/mpi4py/MPI.src/Comm.pyx", line 2544, in mpi4py.MPI.Intracomm.Spawn
    with nogil: CHKERR( MPI_Comm_spawn(
mpi4py.MPI.Exception: MPI_ERR_INTERN: internal error
[sbuild:00243] 1 more process has sent help message help-mca-bml-r2.txt / unreachable proc
[sbuild:00243] 1 more process has sent help message help-mpi-runtime.txt / mpi_init:startup:internal-failure
[sbuild:00243] 1 more process has sent help message help-mpi-errors.txt / mpi_errors_are_fatal unknown handle
E: Build killed with signal TERM after 150 minutes of inactivity

The error is intermittent, indicating a race condition (may pass
after trying the build again).

Upstream acknowledges spawn is flakey in OpenMPI 4 (likely fixed in
OpenMPI 5), but the suggested workaround MPI4PY_TEST_SPAWN=false does
not prevent spawn tests from launching.



-- System Information:
Debian Release: trixie/sid
  APT prefers unstable-debug
  APT policy: (500, 'unstable-debug'), (500, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 6.10.6-amd64 (SMP w/8 CPU threads; PREEMPT)
Locale: LANG=en_AU.UTF-8, LC_CTYPE=en_AU.UTF-8 (charmap=UTF-8), LANGUAGE=en_AU:en
Shell: /bin/sh linked to /usr/bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages python3-mpi4py depends on:
ii  libc6            2.40-2
ii  libopenmpi3t64   4.1.6-13.3
ii  mpi-default-bin  1.17
ii  python3          3.12.5-1

python3-mpi4py recommends no packages.

Versions of packages python3-mpi4py suggests:
ii  python3-numpy  1:1.26.4+ds-11

-- no debconf information



More information about the debian-science-maintainers mailing list