[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[netCDF #SZV-506829]: excessive warnings when accessing HDF4 from netCDF



Hi Charlie,

> Here is a NASA TRMM dataset that demonstrates the issue
> 
> http://dust.ess.uci.edu/hdf/3B43.070901.6A.HDF
> 
> ncks -D dbg_lvl now calls nc_set_log_level(dbg_lvl).
> The HDF5 error for "file signature" is present and logged for all
> values of dbg_lvl >= 0 on this file.
> Interestingly, using dbg_lvl=5 does not increase the verbosity of
> logging at all for this particular command on this dataset.
> So the error logging seems verbose and insensitive, or this particular
> HDF5 diagnostic is considered an important error, or both.
> Some HDF4 files (produced with more recent HDF4?) do not trigger the
> "file signature" HDF5 diagnostic, though most NASA HDF4 datasets that
> I have tried do.
> Thus I stand by my suggestion to reduce the frequency or severity of
> logging this particular error, given that it is widely expected and
> thus potentially obfuscates other more important diagnostics.

Thanks for persisting in reporting this issue.  Here's a Jira ticket I created 
for it, in case you want to follow any progress:

   https://bugtracking.unidata.ucar.edu/browse/NCF-306

--Russ

> zender@givre:~$ ncks -O -D 0 ${DATA}/hdf/3B43.070901.6A.HDF ~/foo.nc
> HDF5-DIAG: Error detected in HDF5 (1.8.11) thread 0:
> #000: ../../src/H5Fsuper.c line 155 in H5F_locate_signature(): unable
> to find a valid file signature
> major: Low-level I/O
> minor: Unable to initialize object
> HDF5-DIAG: Error detected in HDF5 (1.8.11) thread 0:
> #000: ../../src/H5Fsuper.c line 155 in H5F_locate_signature(): unable
> to find a valid file signature
> major: Low-level I/O
> minor: Unable to initialize object
> zender@givre:~$
> 
> On 06/05/2014 01:15 PM, Unidata netCDF Support wrote:
> > Hi Charlie,
> >
> >> netCDF configured with --enable-logging reports many (!) of the
> >> warnings when NCO uses it to examine HDF4 files:
> >> "H5F_locate_signature(): unable to find a valid file signature"
> >> Is there a way to quiet this particular warning?
> >> It is a predictable message that could perhaps be fixed, e.g., teach
> >> netCDF4/HDF5 to expect that valid HDF4 signatures are unavailable.
> >> Or perhaps there is a way to set the verbosity of HDF logging?
> >>
> >> Configuring without --enable-logging eliminates the unwanted messages.
> >> But that would also eliminate the unpredictable and wanted messages.
> >
> > If you build netCDF with --enable-logging, then put the following in
> > your code:
> >
> > nc_set_log_level(3);
> >
> > You will then get a ton of output. Trying changing the "3" to a "1" to
> > get less output, or to a 5 to get more.
> >
> >> Something for Unidata and THG to chew-on, I suppose.
> >
> > Yes, we're glad THG provided this, so we could make it available in the
> > netCDF C and Fortran APIs.
> >
> > --Russ
> >
> > Russ Rew                                         UCAR Unidata Program
> > address@hidden                      http://www.unidata.ucar.edu
> >
> >
> >
> > Ticket Details
> > ===================
> > Ticket ID: SZV-506829
> > Department: Support netCDF
> > Priority: Normal
> > Status: Closed
> >
> 
> --
> Charlie Zender, Earth System Sci. & Computer Sci.
> University of California, Irvine 949-891-2429 )'(
> 
> 
Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu



Ticket Details
===================
Ticket ID: SZV-506829
Department: Support netCDF
Priority: Normal
Status: Closed