> On 26 Jan 2015, at 12:36, Jan Blechta <[email protected]> wrote: > > On Mon, 26 Jan 2015 12:11:18 +0100 > Mikael Mortensen <[email protected]> wrote: > >> >>> On 26 Jan 2015, at 11:53, Jan Blechta <[email protected]> >>> wrote: >>> >>> On Mon, 26 Jan 2015 11:42:42 +0100 >>> Mikael Mortensen <[email protected] >>> <mailto:[email protected]>> wrote: >>> >>>> Hi, >>>> >>>> With recent (today) hashdist installation I run the following >>>> test.py script >>>> >>>> from dolfin import * >>>> >>>> mesh = UnitSquareMesh(4, 4) >>>> V = VectorFunctionSpace(mesh, 'CG', 1) >>>> u = interpolate(Expression(('x[0]', 'x[1]')), V) >>>> print “Hi" >>>> >>>> >>>> [mikael@ubuntu tests (master)]$ python test.py >>>> Hi >>>> >>>> and then it just hangs there without exiting. Same thing in >>>> ipython. I am currently using host python, but I got the same when >>>> I compiled with hashdist python. Anyone else experiencing this >>>> weirdness? Any idea what is going on? I’m on Ubuntu 14.04 and I’ve >>>> done instant-clean. >>> >>> To debug it, you could try attaching debugger (if it is allowed on >>> your system) >>> >>> gdb python <pid> >>> >>> hit ^C, and print stacktrace >>> >>> (gdb) bt >>> >>> Jan >> >> Thanks for the tip. Not really familiar with gdb so I tried pdb first >> >> python -m pdb test_GaussDivergence.py >>> /home/mikael/MySoftware/fenicstools/tests/test.py(5)<module>() >> -> from dolfin import * >> (Pdb) continue >> Hi >> The program finished and will be restarted >> >> and then it just hangs there. Ctrl+C does nothing. >> >> >> With gdb I am getting somewhere though: >> >> gdb python >> Reading symbols from python...(no debugging symbols found)...done. >> (gdb) run test.py >> Starting >> program: /home/mikael/.hashdist/bld/profile/6axink4nal3d/bin/python >> test.py process 16999 is executing new >> program: /home/mikael/.hashdist/bld/python/qlvjlzdbishm/bin/python2.7 >> [Thread debugging using libthread_db enabled] Using host libthread_db >> library "/lib/x86_64-linux-gnu/libthread_db.so.1". [New Thread >> 0x7fffdc7e3700 (LWP 17005)] Hi ^C >> Program received signal SIGINT, Interrupt. >> __lll_lock_wait () >> at ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:135 >> 135 ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S: No >> such file or directory. (gdb) bt #0 __lll_lock_wait () >> at ../nptl/sysdeps/unix/sysv/linux/x86_64/lowlevellock.S:135 #1 >> 0x00007ffff7bc6657 in _L_lock_909 () >> from /lib/x86_64-linux-gnu/libpthread.so.0 #2 0x00007ffff7bc6480 in >> __GI___pthread_mutex_lock (mutex=0x7ffff063fa10 <attribute_lock+16>) >> at ../nptl/pthread_mutex_lock.c:79 #3 0x00007ffff03a140c in >> ompi_attr_get_c () >> from /home/mikael/.hashdist/bld/mpi/vrn4awnke2cx/lib/libmpi.so.1 #4 >> 0x00007ffff03cba27 in MPI_Attr_get () >> from /home/mikael/.hashdist/bld/mpi/vrn4awnke2cx/lib/libmpi.so.1 #5 >> 0x00007ffff07b2afb in Petsc_DelComm_Outer () >> from /home/mikael/.hashdist/bld/petsc/dfw5fwbfp6hl/lib/libpetsc.so.3.5 >> #6 0x00007ffff03a1e88 in ompi_attr_delete () >> from /home/mikael/.hashdist/bld/mpi/vrn4awnke2cx/lib/libmpi.so.1 #7 >> 0x00007ffff03cb93c in MPI_Attr_delete () >> from /home/mikael/.hashdist/bld/mpi/vrn4awnke2cx/lib/libmpi.so.1 #8 >> 0x00007ffff07a2baa in PetscCommDestroy () >> from /home/mikael/.hashdist/bld/petsc/dfw5fwbfp6hl/lib/libpetsc.so.3.5 >> #9 0x00007ffff07a52bc in PetscHeaderDestroy_Private () >> from /home/mikael/.hashdist/bld/petsc/dfw5fwbfp6hl/lib/libpetsc.so.3.5 >> #10 0x00007ffff081372c in ISLocalToGlobalMappingDestroy () >> from /home/mikael/.hashdist/bld/petsc/dfw5fwbfp6hl/lib/libpetsc.so.3.5 >> #11 0x00007ffff0819d08 in PetscLayoutDestroy () >> from /home/mikael/.hashdist/bld/petsc/dfw5fwbfp6hl/lib/libpetsc.so.3.5 >> #12 0x00007ffff0871718 in VecDestroy () >> from /home/mikael/.hashdist/bld/petsc/dfw5fwbfp6hl/lib/libpetsc.so.3.5 >> #13 0x00007ffff22bb15c in dolfin::PETScVector::~PETScVector() () >> from /home/mikael/.hashdist/bld/dolfin/wt5zll4ojwwy/lib/libdolfin.so.1.5 >> #14 0x00007ffff22bb229 in dolfin::PETScVector::~PETScVector() () >> from /home/mikael/.hashdist/bld/dolfin/wt5zll4ojwwy/lib/libdolfin.so.1.5 >> #15 0x00007ffff221c34a in dolfin::Function::~Function() () >> from /home/mikael/.hashdist/bld/dolfin/wt5zll4ojwwy/lib/libdolfin.so.1.5 >> #16 0x00007ffff221c4d9 in dolfin::Function::~Function() () >> from /home/mikael/.hashdist/bld/dolfin/wt5zll4ojwwy/lib/libdolfin.so.1.5 >> #17 0x00007ffff308f999 in >> std::_Sp_counted_base<(__gnu_cxx::_Lock_policy)2>::_M_release >> (this=0x19bbc60) at /usr/include/c++/4.8/bits/shared_ptr_base.h:144 >> #18 0x00007fffdf8093cc in _wrap_delete_Function () >> from >> /home/mikael/Software/hashstack/default/lib/python2.7/site-packages/dolfin/cpp/_function.so >> #19 0x00007fffdf7fc15e in SwigPyObject_dealloc () >> from >> /home/mikael/Software/hashstack/default/lib/python2.7/site-packages/dolfin/cpp/_function.so >> #20 0x00000000005392ff in ?? () #21 0x00000000004d914b in ?? () #22 >> 0x00000000004fdb96 in PyDict_SetItem () #23 0x000000000055a9e1 in >> _PyModule_Clear () #24 0x00000000004f2ad4 in PyImport_Cleanup () #25 >> 0x000000000042fa89 in Py_Finalize () #26 0x000000000046ac10 in >> Py_Main () > > Can you reproduce it with PETSc example > http://www.mcs.anl.gov/petsc/petsc-current/src/vec/is/examples/tests/ex1.c.html > which also calls ISLocalToGlobalMappingDestroy? > Just navigate to <petsc>/src/vec/is/examples/tests/, type make and > execute. > > Jan >
Problem went away with openmpi 1.6.5. Btw, is the petsc source stored somewhere with hashdist? M >> >> So I’m trying to google this now. I think it’s because I’m using a >> newer version of openmpi than in the default hashdist script. Default >> is 1.6.5, but I have compiled with 1.8.2. 1.6.5 gives me problems >> with h5py, which is why I upgraded it in the first place. So at the >> moment h5py works, but not fenics:-( >> >> M >> >>> >>>> >>>> Mikael >>>> >>>> >>>> _______________________________________________ >>>> fenics-support mailing list >>>> [email protected] >>>> <mailto:[email protected]> >>>> http://fenicsproject.org/mailman/listinfo/fenics-support >>>> <http://fenicsproject.org/mailman/listinfo/fenics-support> > _______________________________________________ fenics-support mailing list [email protected] http://fenicsproject.org/mailman/listinfo/fenics-support
