Skip to main content
Version: Latest

Troubleshooting

When a run stops, DynEarthSol prints a two-digit exit code and its category. The first digit says where to look: 1x is the input, 2x the environment (files, libraries), and 3x–6x the computation or the code itself. The full table is in exit codes.

Building​

SymptomCheck
HDF5 headers or library not foundInstall the development package (libhdf5-dev on Ubuntu, hdf5 with Homebrew). Python's h5py does not provide it. Set HDF5_INCLUDE_DIR and HDF5_LIB_DIR only if the wrong installation is picked up. Or build without it: make hdf5=0.
omp.h or libomp not found on macOSbrew install libomp, then rebuild.
Missing nanoflann, knn-bvh or mmg sourcesThe repository uses submodules: clone with --recursive, or run git submodule update --init.
An option a page uses is not recognizedThe executable is from an older revision. Check ./dynearthsol2d -h and the revision a tutorial was checked with (its At a glance box).

See Installation for dependencies and build options.

Starting a run​

SymptomCheck
Exit 10 or 11 before the first stepA configuration error: a misspelled or retired parameter, or a value out of range. The message names the parameter. Look it up in the parameter reference.
Exit 12The .poly or .exo mesh file is malformed. See making a mesh with a .poly file.
.poly file not foundRun from the directory that holds both the input file and the .poly file, or give its path in mesh.poly_filename.
Exit 21 or 22 when restartingThe checkpoint does not match this executable: it was written by an older file-format revision or lacks records this version needs. See checkpointing and restart.
Exit 30 or 31The feature is not available in this dimension, or the executable was built without the library it needs (for example MMG or GoSPL).

During a run​

SymptomCheck
Exit 50: NaN or non-finite valueThe run prints the affected fields and the first bad element and node. Common causes are a time step too large (control.dt_fraction), a control.inertial_scaling so small that inertia dominates (see mass scaling), or a badly distorted mesh.
Exit 54: PT did not convergeSee the PT controls in choosing a mechanical solver.
Exit 40–42: meshing failedThe mesh became too distorted for Triangle, TetGen or MMG to rebuild. Check the quality settings in [mesh] and the remeshing options.
Warning that element groups may share nodesA possible OpenMP race. Increase the resolution or use fewer threads.
Very slow on a small meshToo many threads for the work: set OMP_NUM_THREADS=1 (or a few) and OMP_DYNAMIC=FALSE.

Output​

SymptomCheck
No .vtkhdf filesVTKHDF needs an HDF5 build (hdf5=1). Otherwise convert the binary frames with 2vtk.py.
2vtk.py skips frames of a restarted runConvert with 2vtk.py -g <restarted modelname> to include the parent run's frames. See visualizing outputs.
Lost .info fileutils/recreate_info.py rebuilds it from the frame files.

Getting help​

If the problem remains, open an issue in the DynEarthSol issue tracker with the input file, the exit code and message, and the run's <modelname>.manifest, which records how the executable was built and run.