[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#933844: libpmi-pmix-dev: MPI failure in dolfin tests



Package: libpmi-pmix-dev
Version: 3.1.4~rc2-1
Severity: serious
Justification: breaks MPI client tests

Hi Alistair and Debian science,

I'm upgrading dolfin/fenics to 2019.1.0 alongside HYPRE 2.16.0, PETSc
3.11.3, pybind11 2.3.0.  The tests of the libraries have passed, but
the build is failing badly with an MPI error:

  Run C++ regressions tests (serial)
  Test project /home/projects/fenics/build/dolfin/obj-x86_64-linux-gnu
  ...
        Start  13: demo_eigenvalue_serial
   3/51 Test  #13: demo_eigenvalue_serial ....................***Failed    0.02 sec
  *** The MPI_Comm_rank() function was called before MPI_INIT was invoked.
  *** This is disallowed by the MPI standard.
  *** Your MPI job will now abort.
  [grendel:16518] Local abort before MPI_INIT completed completed successfully, but am not able to aggregate error messages, and not able to guarantee that all other processes were killed!

The same error is also triggered when running with mpirun, so the
problem is not due to running an MPI-linked program as a serial job.

I tested the build last week before proceeding with the upgrades and
package uploads, all was fine then. The nature and the timing of the
error suggests to me that pmix 3.1.4~rc2-1 (built last Thursday
1/8/2019) might be involved.  

It doesn't make sense that pmix should trigger an error like this, but
I've filed this RC bug against pmix to halt migration while we look
into it.

I've raised a discussion thread with FEniCS upstream at
https://fenicsproject.slack.com/archives/C26N589GV/p1564913720002300

Drew


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

Kernel: Linux 4.19.0-5-amd64 (SMP w/4 CPU cores)
Kernel taint flags: TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=en_AU.UTF-8, LC_CTYPE=en_AU.UTF-8 (charmap=UTF-8), LANGUAGE=en_AU.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled

Versions of packages libpmi-pmix-dev depends on:
ii  libpmi1-pmix  3.1.4~rc2-1
ii  libpmi2-pmix  3.1.4~rc2-1

libpmi-pmix-dev recommends no packages.

libpmi-pmix-dev suggests no packages.

-- no debconf information


Reply to: