Open sadielbartholomew opened 1 year ago
I wonder if #531 has fixed this?
Thanks for the heads up David, I'll look into it once I've tackled higher-priority work...
Still seeing this now with the current main
branch. Specifically, running test_Field.py
shows (highly repeatedly) the following across multiple test methods, including the initial one, test_Field__add__
:
HDF5-DIAG: Error detected in HDF5 (1.14.0) thread 1:
#000: H5A.c line 679 in H5Aopen_by_name(): unable to synchronously open attribute
major: Attribute
minor: Can't open object
#001: H5A.c line 641 in H5A__open_by_name_api_common(): unable to open attribute: '_QuantizeBitRoundNumberOfSignificantBits'
major: Attribute
minor: Can't open object
#002: H5A.c line 464 in H5A__open_common(): unable to open attribute: '_QuantizeBitRoundNumberOfSignificantBits'
major: Attribute
minor: Can't open object
#003: H5VLcallback.c line 1138 in H5VL_attr_open(): attribute open failed
major: Virtual Object Layer
minor: Can't open object
#004: H5VLcallback.c line 1105 in H5VL__attr_open(): attribute open failed
major: Virtual Object Layer
minor: Can't open object
#005: H5VLnative_attr.c line 161 in H5VL__native_attr_open(): can't open attribute
major: Attribute
minor: Can't open object
#006: H5Aint.c line 658 in H5A__open_by_name(): unable to load attribute info from object header
major: Attribute
minor: Unable to initialize object
#007: H5Oattribute.c line 502 in H5O__attr_open_by_name(): can't locate attribute: '_QuantizeBitRoundNumberOfSignificantBits'
major: Attribute
minor: Object not found
Which version of netcdf4-python are you using?
I am using 1.6.3
, specifically with:
$ conda list | grep netcdf
libnetcdf 4.9.2 nompi_hf3f8848_103 conda-forge
netcdf-flattener 1.2.0 pyh9f0ad1d_0 conda-forge
netcdf-fortran 4.6.0 nompi_h3142581_105 conda-forge
netcdf4 1.6.3 nompi_py38hd31e641_102 conda-forge
which appears to be the latest version having checked over on https://github.com/Unidata/netcdf4-python. What version are you using and are you seeing this kind of message crop up?
Some CI jobs have produced output indicating errors in HDF5 threads when
test_Data_get_filenames
runs: see report in https://github.com/NCAS-CMS/cf-python/pull/520#issuecomment-1350244307 (with further detail in https://github.com/NCAS-CMS/cf-python/pull/520#issuecomment-1351305513). As deduced by @davidhassell in https://github.com/NCAS-CMS/cf-python/pull/520#issuecomment-1350649482, this is probably not an issue with the cf-python code, rather HDF5 and netCDF-C, but I'm opening this Issue because I/we don't have time to look into it now and we should confirm it is indeed 'not our problem' (and find a way to prevent the errors showing up regardless).