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

[netCDF #DRB-862049]: Build of netCDF-4 on x86_64 system?

> 
> Hi guys,
> 
> I've been building szip-2.0, HDF5-1.8.0-beta1, and netCDF-4.0-beta1
> on various systems to get support for netCDF 4 in our NCL and
> Python code.
> 
> I've been successful on 32-bit systems, but the "make check"
> failed on an x86_64 Linux system, using gcc and g++.
> 
> I don't know if you are interested in the output from this build,
> but I've attached it to this file in case you're interested.
> 
> Also, the "configure" script failed because you need to link "-lsz"
> when you link "-lhdf5" or "-lhdf5_hl". Rick Grubin gave me a fix
> to work around this problem.
> 
> --Mary
> 

Howdy Mary!

I know about the -lsz thing, I'll fix it in the snapshot soon.

According to your output, something is wrong with the HDF5 zlib compression:

*** Checking HDF5 deflate filter setting and getting...Sorry! Unexpected 
result, tst_h_vars.c, line: 228
HDF5-DIAG: Error detected in HDF5 (1.8.0-beta1) thread 0:
  #000: H5Z.c line 1270 in H5Zget_filter_info(): Filter not defined
    major: Invalid arguments to routine
    minor: Bad value
  #001: H5Z.c line 914 in H5Z_find(): required filter is not registered
    major: Data filters
    minor: Object not found
1 failures
*** Checking HDF5 deflatfe, fletcher32, shuffle filter setting and 
getting...Sorry! Unexpected result, tst_h_vars.c, line: 300
1 failures

This is in a tst_h_* program, which all check the HDF5 layer without involving 
any netCDF-4 code.

How was your HDF5 build? Did you use the --with-zlib= option?

Perhaps it was your -lsz fix which caused the proper zlib library to be missing.

I will let you know when I get a sz library up and running with netCDF-4 here...

Thanks,

Ed

Ticket Details
===================
Ticket ID: DRB-862049
Department: Support netCDF
Priority: High
Status: Closed