ADL_Unpacker with ADL4/Mx6.2 crashing AFTER completion

[houchin]

Hi all,

I'm beta testing a tarball release of ADL 4 with MX6.2, and I'm getting HDF related crashes from the ADL_Unpacker.

In my environment, I have the COTS compiled per the UW script, and have as distributed ADL 4 built against that. That runs just fine.

I have the tarball release, with only a very minor change against the imakeconf files mimicing what's in the UW modifications for ADL 4, and it compiles fine. I am linking against the very same compiled COTS directory as works just fine with the UW distributed ADL 4.

The crash comes AFTER the unpacker finishes; all of the binary and .asc files are there, and in the log it states that it's finished storing the information for the RDR file I'm unpacking. It appears to be crashing in some cleanup phase after the ADL_Unpacker main exits (see backtrace below).

Any ideas? I searched for "H5Eset_auto1" in the ADL code, and I didn't find it.

2012/08/22 21:25:38.033.504 (14130.47493257320656): DBG_HIGH DDSASC_Granule.cpp|344|1345670738::DDSASC_Granule.cpp::storeGranuleToDms():344 - Normal Granule '/Data_Products/VIIRS-SCIENCE-RDR/VIIRS-SCIENCE-RDR_Gran_0' from HDF has been stored in DMS with data and metadata, with the following URID: 50354e4f-a893d-a8c0f809-14b1ede5
Collection Short Name: VIIRS-SCIENCE-RDR

2012/08/22 21:25:38.033.550 (14130.47493257320656): DBG_HIGH DDSASC_Granule.cpp|363|1345670738::DDSASC_Granule.cpp::storeGranuleToDms():363 - Finished storing in DMS the granule information for /Data_Products/VIIRS-SCIENCE-RDR/VIIRS-SCIENCE-RDR_Gran_0


2012/08/22 21:25:38.212.996 (14130.47493257320656): DBG_HIGH DDSASC_H5FileProperties.cpp|607|1345670738::DDSASC_H5FileProperties.cpp::myHDF5ErrorHandler():607 - HDF5 error:
HDF5-DIAG: Error detected in HDF5 (1.8.7) thread 0:
#000: H5E.c line 1592 in H5Eget_auto2(): wrong API function, H5Eset_auto1 has been called
major: Error API
minor: Can't get value[/code]

Code: Select all

#0  0x0000003c55030285 in raise () from /lib64/libc.so.6
(gdb) where
#0  0x0000003c55030285 in raise () from /lib64/libc.so.6
#1  0x0000003c55031d30 in abort () from /lib64/libc.so.6
#2  0x0000000000476850 in myHDF5ErrorHandler (clientData=0x0) at DDSASC_H5FileProperties.cpp:616
#3  0x00002b31d9d2b9c1 in H5E_dump_api_stack () from /daq3_raid3/user/aero/aeroadl/latest/common/local/lib/libhdf5.so.7
#4  0x00002b31d9d289c8 in H5Eget_auto2 () from /daq3_raid3/user/aero/aeroadl/latest/common/local/lib/libhdf5.so.7
#5  0x00002b31d9cc7a78 in H5_term_library () from /daq3_raid3/user/aero/aeroadl/latest/common/local/lib/libhdf5.so.7
#6  0x0000003c55033515 in exit () from /lib64/libc.so.6
#7  0x0000003c5501d99b in __libc_start_main () from /lib64/libc.so.6
#8  0x0000000000418049 in _start ()

[houchin]

Are you sure this software you install completely?

Yes. I've done further dumping of the environment and rebuilding, and best I can tell absolutely everything is identical in terms of COTS dependencies between my ADL 4 version (which works) and the ADL Mx6.2 version (tarball). I've even run ldd at the suggestion of Raytheon, and it's reporting that both versions of the ADL_Unpacker are running against the same version of the HDF libraries.

Plus, the ADL VIIRS Controller does NOT crash in either version. Something really weird is happening...