> 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

Reply via email to