I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ? I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag? HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs and finally, mbconvert: mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs But when I do that, I get this error: HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp ==rob
Rob, We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ? We could try to replicate your configuration and push a fix if there is a problem. Vijay On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
On 06/09/2016 01:18 PM, Vijay S. Mahadevan wrote:
Rob,
We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ?
We could try to replicate your configuration and push a fix if there is a problem.
I'm happy to include the config.log. It's attached to this email. An identically-configured HDF5-1.8.17 works well. What version of HDF5 created the unit test files? git tells me they were committed in July 2013, but HDF5 libraries are supposed to be backwards compatible. ==rob
Vijay
On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
Ok, looks like configuration passed cleanly. And certainly looked like your linkage didn't have problems either. Can you try running make check in interactive mode to see if that passes in test/io build folder ? If these fail, then it may point to the root cause. Either way, I'll try to build MOAB and point to your HDF5 version to replicate issues locally. More soon. Vijay On Thu, Jun 9, 2016 at 1:49 PM, Rob Latham <[email protected]> wrote:
On 06/09/2016 01:18 PM, Vijay S. Mahadevan wrote:
Rob,
We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ?
We could try to replicate your configuration and push a fix if there is a problem.
I'm happy to include the config.log. It's attached to this email.
An identically-configured HDF5-1.8.17 works well.
What version of HDF5 created the unit test files? git tells me they were committed in July 2013, but HDF5 libraries are supposed to be backwards compatible.
==rob
Vijay
On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
On 06/09/2016 01:57 PM, Vijay S. Mahadevan wrote:
Ok, looks like configuration passed cleanly. And certainly looked like your linkage didn't have problems either. Can you try running make check in interactive mode to see if that passes in test/io build folder ? If these fail, then it may point to the root cause.
all red across the board! ../../../config/test-driver: line 107: 327 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: readutil_test ../../../config/test-driver: line 107: 334 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: tqdcfr ../../../config/test-driver: line 107: 342 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: vtk_test FAIL: cub_file_test FAIL: gmsh_test FAIL: nastran_test FAIL: ideas_test FAIL: stl_test FAIL: smf_test ============================================================================ Testsuite summary for MOAB 4.9.2pre ============================================================================ # TOTAL: 9 # PASS: 0 # SKIP: 0 # XFAIL: 0 # FAIL: 9 # XPASS: 0 # ERROR: 0 ============================================================================ See test/io/test-suite.log Please report to [email protected] ============================================================================ make[2]: *** [test-suite.log] Error 1 make[2]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make[1]: *** [check-TESTS] Error 2 make[1]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make: *** [check-am] Error 2 I can stick with hdf5-1.8.17 for now. ==rob
Either way, I'll try to build MOAB and point to your HDF5 version to replicate issues locally. More soon.
Vijay
On Thu, Jun 9, 2016 at 1:49 PM, Rob Latham <[email protected]> wrote:
On 06/09/2016 01:18 PM, Vijay S. Mahadevan wrote:
Rob,
We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ?
We could try to replicate your configuration and push a fix if there is a problem.
I'm happy to include the config.log. It's attached to this email.
An identically-configured HDF5-1.8.17 works well.
What version of HDF5 created the unit test files? git tells me they were committed in July 2013, but HDF5 libraries are supposed to be backwards compatible.
==rob
Vijay
On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
all red across the board!
Hmm that's bad. We are checking now.
I can stick with hdf5-1.8.17 for now.
Sounds good. I'll update when we have fixed the issue. Vijay On Thu, Jun 9, 2016 at 2:20 PM, Rob Latham <[email protected]> wrote:
On 06/09/2016 01:57 PM, Vijay S. Mahadevan wrote:
Ok, looks like configuration passed cleanly. And certainly looked like your linkage didn't have problems either. Can you try running make check in interactive mode to see if that passes in test/io build folder ? If these fail, then it may point to the root cause.
all red across the board!
../../../config/test-driver: line 107: 327 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: readutil_test ../../../config/test-driver: line 107: 334 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: tqdcfr ../../../config/test-driver: line 107: 342 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: vtk_test FAIL: cub_file_test FAIL: gmsh_test FAIL: nastran_test FAIL: ideas_test FAIL: stl_test FAIL: smf_test ============================================================================ Testsuite summary for MOAB 4.9.2pre ============================================================================ # TOTAL: 9 # PASS: 0 # SKIP: 0 # XFAIL: 0 # FAIL: 9 # XPASS: 0 # ERROR: 0 ============================================================================ See test/io/test-suite.log Please report to [email protected] ============================================================================ make[2]: *** [test-suite.log] Error 1 make[2]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make[1]: *** [check-TESTS] Error 2 make[1]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make: *** [check-am] Error 2
I can stick with hdf5-1.8.17 for now.
==rob
Either way, I'll try to build MOAB and point to your HDF5 version to replicate issues locally. More soon.
Vijay
On Thu, Jun 9, 2016 at 1:49 PM, Rob Latham <[email protected]> wrote:
On 06/09/2016 01:18 PM, Vijay S. Mahadevan wrote:
Rob,
We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ?
We could try to replicate your configuration and push a fix if there is a problem.
I'm happy to include the config.log. It's attached to this email.
An identically-configured HDF5-1.8.17 works well.
What version of HDF5 created the unit test files? git tells me they were committed in July 2013, but HDF5 libraries are supposed to be backwards compatible.
==rob
Vijay
On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
Hi Rob, I think there is some incompatibility with mpi version /homes/robl/soft/mpich-logfs/bin/mpicxx does not work ________________________________________ From: [email protected] [[email protected]] on behalf of Rob Latham [[email protected]] Sent: Thursday, June 09, 2016 2:20 PM To: Vijay S. Mahadevan Cc: [email protected] Subject: Re: [MOAB-dev] mbconvert and HDF5-1.10 On 06/09/2016 01:57 PM, Vijay S. Mahadevan wrote:
Ok, looks like configuration passed cleanly. And certainly looked like your linkage didn't have problems either. Can you try running make check in interactive mode to see if that passes in test/io build folder ? If these fail, then it may point to the root cause.
all red across the board! ../../../config/test-driver: line 107: 327 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: readutil_test ../../../config/test-driver: line 107: 334 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: tqdcfr ../../../config/test-driver: line 107: 342 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: vtk_test FAIL: cub_file_test FAIL: gmsh_test FAIL: nastran_test FAIL: ideas_test FAIL: stl_test FAIL: smf_test ============================================================================ Testsuite summary for MOAB 4.9.2pre ============================================================================ # TOTAL: 9 # PASS: 0 # SKIP: 0 # XFAIL: 0 # FAIL: 9 # XPASS: 0 # ERROR: 0 ============================================================================ See test/io/test-suite.log Please report to [email protected] ============================================================================ make[2]: *** [test-suite.log] Error 1 make[2]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make[1]: *** [check-TESTS] Error 2 make[1]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make: *** [check-am] Error 2 I can stick with hdf5-1.8.17 for now. ==rob
Either way, I'll try to build MOAB and point to your HDF5 version to replicate issues locally. More soon.
Vijay
On Thu, Jun 9, 2016 at 1:49 PM, Rob Latham <[email protected]> wrote:
On 06/09/2016 01:18 PM, Vijay S. Mahadevan wrote:
Rob,
We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ?
We could try to replicate your configuration and push a fix if there is a problem.
I'm happy to include the config.log. It's attached to this email.
An identically-configured HDF5-1.8.17 works well.
What version of HDF5 created the unit test files? git tells me they were committed in July 2013, but HDF5 libraries are supposed to be backwards compatible.
==rob
Vijay
On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
Rob, all tests work fine with the following options: ../configure CC=mpicc CXX=mpicxx --with-hdf5=/homes/robl/soft/hdf5-1.10.0-mpich-logfs --with-mpi --disable-fortran So I agree with Iulian that your custom MPI installation is messed up. If you pass CC/CXX, make sure they are consistent with the "--with-mpi" option. Vijay On Thu, Jun 9, 2016 at 2:45 PM, Grindeanu, Iulian R. <[email protected]> wrote:
Hi Rob, I think there is some incompatibility with mpi version
/homes/robl/soft/mpich-logfs/bin/mpicxx
does not work ________________________________________ From: [email protected] [[email protected]] on behalf of Rob Latham [[email protected]] Sent: Thursday, June 09, 2016 2:20 PM To: Vijay S. Mahadevan Cc: [email protected] Subject: Re: [MOAB-dev] mbconvert and HDF5-1.10
On 06/09/2016 01:57 PM, Vijay S. Mahadevan wrote:
Ok, looks like configuration passed cleanly. And certainly looked like your linkage didn't have problems either. Can you try running make check in interactive mode to see if that passes in test/io build folder ? If these fail, then it may point to the root cause.
all red across the board!
../../../config/test-driver: line 107: 327 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: readutil_test ../../../config/test-driver: line 107: 334 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: tqdcfr ../../../config/test-driver: line 107: 342 Segmentation fault (core dumped) "$@" > $log_file 2>&1 FAIL: vtk_test FAIL: cub_file_test FAIL: gmsh_test FAIL: nastran_test FAIL: ideas_test FAIL: stl_test FAIL: smf_test ============================================================================ Testsuite summary for MOAB 4.9.2pre ============================================================================ # TOTAL: 9 # PASS: 0 # SKIP: 0 # XFAIL: 0 # FAIL: 9 # XPASS: 0 # ERROR: 0 ============================================================================ See test/io/test-suite.log Please report to [email protected] ============================================================================ make[2]: *** [test-suite.log] Error 1 make[2]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make[1]: *** [check-TESTS] Error 2 make[1]: Leaving directory `/blues/gpfs/home/robl/src/moab.git/build/test/io' make: *** [check-am] Error 2
I can stick with hdf5-1.8.17 for now.
==rob
Either way, I'll try to build MOAB and point to your HDF5 version to replicate issues locally. More soon.
Vijay
On Thu, Jun 9, 2016 at 1:49 PM, Rob Latham <[email protected]> wrote:
On 06/09/2016 01:18 PM, Vijay S. Mahadevan wrote:
Rob,
We haven't tested with the latest HDF5 version yet. We are planning to do this soon. So it is possible that the configuration is broken or that parallel HDF5 was not enabled. Can you send the config.log file so that we can decipher if the configuration was successful to start with ?
We could try to replicate your configuration and push a fix if there is a problem.
I'm happy to include the config.log. It's attached to this email.
An identically-configured HDF5-1.8.17 works well.
What version of HDF5 created the unit test files? git tells me they were committed in July 2013, but HDF5 libraries are supposed to be backwards compatible.
==rob
Vijay
On Thu, Jun 9, 2016 at 1:14 PM, Rob Latham <[email protected]> wrote:
I tried running mbconvert with the recently released HDF5-1.10 library. That library seems unable to read MeshFiles/unittest/64bricks_12ktet.h5m ?
I built my own HDF5-1.10, so maybe I needed to configure with an extra backwards compatibility flag?
HDF5: ../configure --prefix=/homes/robl/soft/hdf5-1.10.0 CC=mpicc \ --enable-parallel
MOAB: ../configure CC=mpicc CXX=mpicxx \ --with-hdf5=/homes/robl/soft/hdf5-1.10.0 \ --with-mpi=/homes/robl/soft/mpich \ --prefix=/homes/robl/soft/moab-logfs
and finally, mbconvert:
mpiexec -np $((PBS_NNODES*4)) ./build-1.8.17/tools/mbconvert \ -O PARALLEL=READ_PART -O PARTITION=PARALLEL_PARTITION \ -O PARALLEL_RESOLVE_SHARED_ENTS -o PARALLEL=WRITE_PART \ MeshFiles/unittest/64bricks_12ktet.h5m /lcrc/project/radix/robl/moab/out-gpfs
But when I do that, I get this error:
HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 575 in H5Fopen(): can't set access and transfer property lists major: File accessibilty minor: Can't set value #001: ../../src/H5Pint.c line 5506 in H5P_verify_apl_and_dxpl(): not the required access property list major: Property lists minor: Inappropriate type #002: ../../src/H5Pint.c line 3767 in H5P_isa_class(): not a property list major: Invalid arguments to routine minor: Inappropriate type HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5G.c line 464 in H5Gopen2(): not a location major: Invalid arguments to routine minor: Inappropriate type #001: ../../src/H5Gloc.c line 253 in H5G_loc(): invalid object ID major: Invalid arguments to routine minor: Bad value HDF5-DIAG: Error detected in HDF5 (1.10.0) MPI-process 0: #000: ../../src/H5F.c line 749 in H5Fclose(): not a file ID major: Invalid arguments to routine minor: Inappropriate type [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: NULL file handle.! [0]MOAB ERROR: is_error() line 133 in src/io/ReadHDF5.hpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed to load file after trying all possible readers! [0]MOAB ERROR: serial_load_file() line 635 in src/Core.cpp [0]MOAB ERROR: --------------------- Error Message ------------------------------------ [0]MOAB ERROR: Failed in step PARALLEL READ PART! [0]MOAB ERROR: load_file() line 553 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 252 in src/parallel/ReadParallel.cpp [0]MOAB ERROR: load_file() line 515 in src/Core.cpp
==rob
participants (3)
-
Grindeanu, Iulian R. -
Rob Latham -
Vijay S. Mahadevan