Hi Brendan,

Could you please try more recent version of zlib? If you see the same problem, 
please send the example that demonstrates the problem to our Helpdesk.

Thank you!

Elena
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Elena Pourmal  The HDF Group  http://hdfgroup.org   
1800 So. Oak St., Suite 203, Champaign IL 61820
217.531.6112
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~




On Jan 27, 2015, at 9:57 AM, Brendan Kochunas <[email protected]> wrote:

> Hi all (new to the list),
> 
> I am currently experiencing/investigating an issue where our usage of HDF5 
> with compression is leading to some defects reported by valgrind (apologies 
> for not having debug symbols for line numbers). These defects are also 
> reported when I run valgrind on
> 
> http://www.hdfgroup.org/ftp/HDF5/examples/examples-by-api/hdf5-examples/1_8/FORTRAN/H5D/h5ex_d_gzip.f90
> 
> The valgrind output I get from the above example is:
> ==54559== Conditional jump or move depends on uninitialised value(s)
> ==54559==    at 0x63664E0: inflateReset2 (in 
> /lib/x86_64-linux-gnu/libz.so.1.2.3.4)
> ==54559==    by 0x63665D8: inflateInit2_ (in 
> /lib/x86_64-linux-gnu/libz.so.1.2.3.4)
> ==54559==    by 0x5989E13: H5Z_filter_deflate (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5.so.8.0.0)
> ==54559==    by 0x59891A0: H5Z_pipeline (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5.so.8.0.0)
> ==54559==    by 0x57C96A1: H5D__chunk_lock (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5.so.8.0.0)
> ==54559==    by 0x57CA5D0: H5D__chunk_read (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5.so.8.0.0)
> ==54559==    by 0x57DA482: H5D__read (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5.so.8.0.0)
> ==54559==    by 0x57DB7B3: H5Dread (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5.so.8.0.0)
> ==54559==    by 0x4E5B143: h5dread_c_ (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5_fortran.so.8.0.0)
> ==54559==    by 0x4E69593: __h5d_provisional_MOD_h5dread_integer_2 (in 
> /opt/hdf5/1.8.11/gcc_4.6.3/lib/libhdf5_fortran.so.8.0.0)
> ==54559==    by 0x40198E: MAIN__ (h5ex_d_gzip.f90:134)
> ==54559==    by 0x401B81: main (h5ex_d_gzip.f90:16)
> 
> In my situation, when I am performing parallel reads I get anomalous 
> behavior, where sometimes the executable works, and sometimes it does not. 
> When it does not work I get the following diagnostic messages from HDF5.
> HDF5-DIAG: Error detected in HDF5 (1.8.13) MPI-process 4223:
>  #000: H5Dio.c line 173 in H5Dread(): can't read data
>    major: Dataset
>    minor: Read failed
>  #001: H5Dio.c line 545 in H5D__read(): can't read data
>    major: Dataset
>    minor: Read failed
>  #002: H5Dchunk.c line 1861 in H5D__chunk_read(): unable to read raw data 
> chunk
>    major: Low-level I/O
>    minor: Read failed
>  #003: H5Dchunk.c line 2891 in H5D__chunk_lock(): data pipeline read failed
>    major: Data filters
>    minor: Filter operation failed
>  #004: H5Z.c line 1383 in H5Z_pipeline(): filter returned failure during read
>    major: Data filters
>    minor: Read failed
>  #005: H5Zdeflate.c line 136 in H5Z_filter_deflate(): memory allocation 
> failed for deflate uncompression
>    major: Resource unavailable
>    minor: No space available for allocation
> 
> Which has a similar call stack, and I wonder if it's trying to allocate using 
> an undefined value. My gut tells me these may be related.
> 
> Is the valgrind issue I am reporting a known issue? Has anyone else run into 
> this?
> 
> Can anyone provide some advice on how might continue debugging this?
> Any thoughts on whether this is an issue with HDF5, zlib or some combination?
> 
> In my use case, it is extremely desirable to use compression as I am dealing 
> with files that are 2.2GB compressed (4.5GB uncompressed). I can probably 
> post some code if needed, but since the example has the same issue perhaps it 
> is better to work from that.
> 
> My hope is that the root cause of this issue is:
> A. Something in my usage of the library
> B. Something wrong with the configure/build/install of the library (which I 
> cannot necessarily change)
> C. Something wrong with the HDF5 library
> 
> Any feedback is welcome.
> 
> Thanks,
> -Brendan
> 
> _______________________________________________
> Hdf-forum is for HDF software users discussion.
> [email protected]
> http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
> Twitter: https://twitter.com/hdf5


_______________________________________________
Hdf-forum is for HDF software users discussion.
[email protected]
http://mail.lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.org
Twitter: https://twitter.com/hdf5

Reply via email to