unsupported data type

Dear All,

when trying to test the previously compiled (with mpicc) HDF5 source I encountered the following error message:

"MPI_ABORT was invoked on rank 3 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.

···

--------------------------------------------------------------------------
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks, combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks, combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks, combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks, combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks, combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks, combiner = 17
--------------------------------------------------------------------------
mpiexec has exited due to process rank 1 with PID 17756 on
node XXXX exiting without calling "finalize". This may
have caused other processes in the application to be
terminated by signals sent by mpiexec (as reported here)."

As I learnt it has to do something with a particular data type (as the error message indicates it), but I'm stuck, I don't know how to proceed.

Could You give some help?

  Best regards,
   Joseph

Hi Joseph,

Could you please indicate the version of the HDF5 library and Open-MPI that you are using?
Furthermore, could you provide a test case that replicates the failure below?

You are correct that it seems that we are constructing a derived datatype that the ADIO driver in ROMIO doesn't seem to be able to parse properly. We don't test however much with Open-MPI, at lease to the extent that we do with MPICH. I'd be curious if you can try the same program with MPICH and see if it gives the same error.

Thanks,
Mohamad

···

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of Kóbori József
Sent: Monday, November 16, 2015 5:30 AM
To: hdf-forum@lists.hdfgroup.org
Subject: [Hdf-forum] unsupported data type

  Dear All,

when trying to test the previously compiled (with mpicc) HDF5 source I
encountered the following error message:

"MPI_ABORT was invoked on rank 3 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on exactly
when Open MPI kills them.
--------------------------------------------------------------------------
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
--------------------------------------------------------------------------
mpiexec has exited due to process rank 1 with PID 17756 on node XXXX
exiting without calling "finalize". This may have caused other processes in the
application to be terminated by signals sent by mpiexec (as reported here)."

As I learnt it has to do something with a particular data type (as the error
message indicates it), but I'm stuck, I don't know how to proceed.

Could You give some help?

  Best regards,
   Joseph

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

Hi Mohamad,

I compiled the latest version of HDF5: hdf5-1.8.16
After compiling (make) I simply typed the "make check" command,
and I got the error below, so I don't have a particular test case.
If You need more info please let me know!
Thank You!

  Regards,
   Joseph

Id�zet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):

···

Hi Joseph,

Could you please indicate the version of the HDF5 library and Open-MPI that you are using?
Furthermore, could you provide a test case that replicates the failure below?

You are correct that it seems that we are constructing a derived datatype that the ADIO driver in ROMIO doesn't seem to be able to parse properly. We don't test however much with Open-MPI, at lease to the extent that we do with MPICH. I'd be curious if you can try the same program with MPICH and see if it gives the same error.

Thanks,
Mohamad

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of K�bori J�zsef
Sent: Monday, November 16, 2015 5:30 AM
To: hdf-forum@lists.hdfgroup.org
Subject: [Hdf-forum] unsupported data type

  Dear All,

when trying to test the previously compiled (with mpicc) HDF5 source I
encountered the following error message:

"MPI_ABORT was invoked on rank 3 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on exactly
when Open MPI kills them.
--------------------------------------------------------------------------
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
--------------------------------------------------------------------------
mpiexec has exited due to process rank 1 with PID 17756 on node XXXX
exiting without calling "finalize". This may have caused other processes in the
application to be terminated by signals sent by mpiexec (as reported here)."

As I learnt it has to do something with a particular data type (as the error
message indicates it), but I'm stuck, I don't know how to proceed.

Could You give some help?

  Best regards,
   Joseph

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

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

-------------------------------------------

J�zsef K�bori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673
E�tv�s Lor�nd University
1117 Budapest, P�zm�ny P. stny. 1/A

Hi Joseph ,

It would be helpful if you would indicate which way you configured HDF5. Using cmake? Using configure? Which options did you use? On which platform?

Helpful also be the content of the file <HDF5_DIR>/lib/libhdf5.settings, where <HDF5_DIR> is the installation directory for HDF5, default "/usr/local" or specified by you using PREFIX or --prefix

Greetings, Richard

···

On 11/24/2015 12:40 PM, K�bori J�zsef wrote:

Hi Mohamad,

I compiled the latest version of HDF5: hdf5-1.8.16
After compiling (make) I simply typed the "make check" command,
and I got the error below, so I don't have a particular test case.
If You need more info please let me know!
Thank You!

Regards,
  Joseph

Id�zet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):

Hi Joseph,

Could you please indicate the version of the HDF5 library and Open-MPI that you are using?
Furthermore, could you provide a test case that replicates the failure below?

You are correct that it seems that we are constructing a derived datatype that the ADIO driver in ROMIO doesn't seem to be able to parse properly. We don't test however much with Open-MPI, at lease to the extent that we do with MPICH. I'd be curious if you can try the same program with MPICH and see if it gives the same error.

Thanks,
Mohamad

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of K�bori J�zsef
Sent: Monday, November 16, 2015 5:30 AM
To: hdf-forum@lists.hdfgroup.org
Subject: [Hdf-forum] unsupported data type

  Dear All,

when trying to test the previously compiled (with mpicc) HDF5 source I
encountered the following error message:

"MPI_ABORT was invoked on rank 3 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on exactly
when Open MPI kills them.
--------------------------------------------------------------------------

Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
--------------------------------------------------------------------------

mpiexec has exited due to process rank 1 with PID 17756 on node XXXX
exiting without calling "finalize". This may have caused other processes in the
application to be terminated by signals sent by mpiexec (as reported here)."

As I learnt it has to do something with a particular data type (as the error
message indicates it), but I'm stuck, I don't know how to proceed.

Could You give some help?

  Best regards,
   Joseph

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

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

-------------------------------------------

J�zsef K�bori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673
E�tv�s Lor�nd University
1117 Budapest, P�zm�ny P. stny. 1/A

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

Hi Richard,

I used "configure" with the following options:
--prefix=/some/dir --enable-parallel

Platform:
Linux atlasz 3.2.0-4-amd64 #1 SMP Debian 3.2.68-1+deb7u1 x86_64 GNU/Linux

The libhdf5.settings file is attached.

Thanks!

  Regards,
   Joseph

Idézet (Richard van Hees <r.m.van.hees@sron.nl>):

libhdf5.settings (3.75 KB)

···

Hi Joseph ,

It would be helpful if you would indicate which way you configured HDF5. Using cmake? Using configure? Which options did you use? On which platform?

Helpful also be the content of the file <HDF5_DIR>/lib/libhdf5.settings, where <HDF5_DIR> is the installation directory for HDF5, default "/usr/local" or specified by you using PREFIX or --prefix

Greetings, Richard

On 11/24/2015 12:40 PM, Kóbori József wrote:

Hi Mohamad,

I compiled the latest version of HDF5: hdf5-1.8.16
After compiling (make) I simply typed the "make check" command,
and I got the error below, so I don't have a particular test case.
If You need more info please let me know!
Thank You!

Regards,
Joseph

Idézet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):

Hi Joseph,

Could you please indicate the version of the HDF5 library and Open-MPI that you are using?
Furthermore, could you provide a test case that replicates the failure below?

You are correct that it seems that we are constructing a derived datatype that the ADIO driver in ROMIO doesn't seem to be able to parse properly. We don't test however much with Open-MPI, at lease to the extent that we do with MPICH. I'd be curious if you can try the same program with MPICH and see if it gives the same error.

Thanks,
Mohamad

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of Kóbori József
Sent: Monday, November 16, 2015 5:30 AM
To: hdf-forum@lists.hdfgroup.org
Subject: [Hdf-forum] unsupported data type

Dear All,

when trying to test the previously compiled (with mpicc) HDF5 source I
encountered the following error message:

"MPI_ABORT was invoked on rank 3 in communicator MPI_COMM_WORLD
with errorcode 1.

NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on exactly
when Open MPI kills them.
-------------------------------------------------------------------------- Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
Error: Unsupported datatype passed to ADIOI_Count_contiguous_blocks,
combiner = 17
-------------------------------------------------------------------------- mpiexec has exited due to process rank 1 with PID 17756 on node XXXX
exiting without calling "finalize". This may have caused other processes in the
application to be terminated by signals sent by mpiexec (as reported here)."

As I learnt it has to do something with a particular data type (as the error
message indicates it), but I'm stuck, I don't know how to proceed.

Could You give some help?

Best regards,
  Joseph

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

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

-------------------------------------------

József Kóbori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673
Eötvös Loránd University
1117 Budapest, Pázmány P. stny. 1/A

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

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

-------------------------------------------

József Kóbori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673
Eötvös Loránd University
1117 Budapest, Pázmány P. stny. 1/A

Hi Joseph,

What is the test that fails in make check? When you run make check, it should tell you which test is running. Please provide that.
I also would like to know the Open-MPI version you are using.

Thanks,
Mohamad

···

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of Kóbori József
Sent: Tuesday, November 24, 2015 8:04 AM
To: hdf-forum@lists.hdfgroup.org
Subject: Re: [Hdf-forum] unsupported data type

  Hi Richard,

I used "configure" with the following options:
--prefix=/some/dir --enable-parallel

Platform:
Linux atlasz 3.2.0-4-amd64 #1 SMP Debian 3.2.68-1+deb7u1 x86_64
GNU/Linux

The libhdf5.settings file is attached.

Thanks!

  Regards,
   Joseph

Idézet (Richard van Hees <r.m.van.hees@sron.nl>):

> Hi Joseph ,
>
> It would be helpful if you would indicate which way you configured
> HDF5. Using cmake? Using configure? Which options did you use? On
> which platform?
>
> Helpful also be the content of the file
> <HDF5_DIR>/lib/libhdf5.settings, where <HDF5_DIR> is the installation
> directory for HDF5, default "/usr/local" or specified by you using
> PREFIX or --prefix
>
> Greetings, Richard
>
> On 11/24/2015 12:40 PM, Kóbori József wrote:
>> Hi Mohamad,
>>
>> I compiled the latest version of HDF5: hdf5-1.8.16 After compiling
>> (make) I simply typed the "make check" command, and I got the error
>> below, so I don't have a particular test case.
>> If You need more info please let me know!
>> Thank You!
>>
>> Regards,
>> Joseph
>>
>> Idézet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):
>>
>>> Hi Joseph,
>>>
>>> Could you please indicate the version of the HDF5 library and
>>> Open-MPI that you are using?
>>> Furthermore, could you provide a test case that replicates the
>>> failure below?
>>>
>>> You are correct that it seems that we are constructing a derived
>>> datatype that the ADIO driver in ROMIO doesn't seem to be able to
>>> parse properly. We don't test however much with Open-MPI, at lease
>>> to the extent that we do with MPICH. I'd be curious if you can try
>>> the same program with MPICH and see if it gives the same error.
>>>
>>> Thanks,
>>> Mohamad
>>>
>>>> -----Original Message-----
>>>> From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On
>>>> Behalf Of Kóbori József
>>>> Sent: Monday, November 16, 2015 5:30 AM
>>>> To: hdf-forum@lists.hdfgroup.org
>>>> Subject: [Hdf-forum] unsupported data type
>>>>
>>>> Dear All,
>>>>
>>>> when trying to test the previously compiled (with mpicc) HDF5
>>>> source I encountered the following error message:
>>>>
>>>> "MPI_ABORT was invoked on rank 3 in communicator
MPI_COMM_WORLD
>>>> with errorcode 1.
>>>>
>>>> NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
>>>> You may or may not see output from other processes, depending on
>>>> exactly when Open MPI kills them.
>>>> -------------------------------------------------------------------
>>>> ------- Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> -------------------------------------------------------------------
>>>> ------- mpiexec has exited due to process rank 1 with PID 17756 on
>>>> node XXXX exiting without calling "finalize". This may have caused
>>>> other processes in the application to be terminated by signals sent
>>>> by mpiexec (as reported here)."
>>>>
>>>> As I learnt it has to do something with a particular data type (as
>>>> the error message indicates it), but I'm stuck, I don't know how to
>>>> proceed.
>>>>
>>>> Could You give some help?
>>>>
>>>> Best regards,
>>>> Joseph
>>>>
>>>>
>>>> _______________________________________________
>>>> Hdf-forum is for HDF software users discussion.
>>>> Hdf-forum@lists.hdfgroup.org
>>>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup
>>>> .org
>>>> Twitter: https://twitter.com/hdf5
>>>
>>> _______________________________________________
>>> Hdf-forum is for HDF software users discussion.
>>> Hdf-forum@lists.hdfgroup.org
>>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.
>>> org
>>> Twitter: https://twitter.com/hdf5
>>>
>>
>>
>>
>> -------------------------------------------
>>
>> József Kóbori
>> PhD Student - Particle Physics and Astronomy Program
>> Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös
>> Loránd University
>> 1117 Budapest, Pázmány P. stny. 1/A
>>
>>
>>
>> _______________________________________________
>> Hdf-forum is for HDF software users discussion.
>> Hdf-forum@lists.hdfgroup.org
>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.o
>> rg
>> Twitter: https://twitter.com/hdf5
>
>
> _______________________________________________
> Hdf-forum is for HDF software users discussion.
> Hdf-forum@lists.hdfgroup.org
> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.or
> g
> Twitter: https://twitter.com/hdf5
>

-------------------------------------------

József Kóbori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös Loránd
University
1117 Budapest, Pázmány P. stny. 1/A

Hi Mohamad,

please found attached the output of "ompi_info" and the log of the faulty hdf5 test.
Thanks!

  Regards,
   Joseph

Idézet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):

hdf5_test.log (4.98 KB)

ompi_info.log (6.68 KB)

···

Hi Joseph,

What is the test that fails in make check? When you run make check, it should tell you which test is running. Please provide that.
I also would like to know the Open-MPI version you are using.

Thanks,
Mohamad

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of Kóbori József
Sent: Tuesday, November 24, 2015 8:04 AM
To: hdf-forum@lists.hdfgroup.org
Subject: Re: [Hdf-forum] unsupported data type

  Hi Richard,

I used "configure" with the following options:
--prefix=/some/dir --enable-parallel

Platform:
Linux atlasz 3.2.0-4-amd64 #1 SMP Debian 3.2.68-1+deb7u1 x86_64
GNU/Linux

The libhdf5.settings file is attached.

Thanks!

  Regards,
   Joseph

Idézet (Richard van Hees <r.m.van.hees@sron.nl>):

> Hi Joseph ,
>
> It would be helpful if you would indicate which way you configured
> HDF5. Using cmake? Using configure? Which options did you use? On
> which platform?
>
> Helpful also be the content of the file
> <HDF5_DIR>/lib/libhdf5.settings, where <HDF5_DIR> is the installation
> directory for HDF5, default "/usr/local" or specified by you using
> PREFIX or --prefix
>
> Greetings, Richard
>
> On 11/24/2015 12:40 PM, Kóbori József wrote:
>> Hi Mohamad,
>>
>> I compiled the latest version of HDF5: hdf5-1.8.16 After compiling
>> (make) I simply typed the "make check" command, and I got the error
>> below, so I don't have a particular test case.
>> If You need more info please let me know!
>> Thank You!
>>
>> Regards,
>> Joseph
>>
>> Idézet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):
>>
>>> Hi Joseph,
>>>
>>> Could you please indicate the version of the HDF5 library and
>>> Open-MPI that you are using?
>>> Furthermore, could you provide a test case that replicates the
>>> failure below?
>>>
>>> You are correct that it seems that we are constructing a derived
>>> datatype that the ADIO driver in ROMIO doesn't seem to be able to
>>> parse properly. We don't test however much with Open-MPI, at lease
>>> to the extent that we do with MPICH. I'd be curious if you can try
>>> the same program with MPICH and see if it gives the same error.
>>>
>>> Thanks,
>>> Mohamad
>>>
>>>> -----Original Message-----
>>>> From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On
>>>> Behalf Of Kóbori József
>>>> Sent: Monday, November 16, 2015 5:30 AM
>>>> To: hdf-forum@lists.hdfgroup.org
>>>> Subject: [Hdf-forum] unsupported data type
>>>>
>>>> Dear All,
>>>>
>>>> when trying to test the previously compiled (with mpicc) HDF5
>>>> source I encountered the following error message:
>>>>
>>>> "MPI_ABORT was invoked on rank 3 in communicator
MPI_COMM_WORLD
>>>> with errorcode 1.
>>>>
>>>> NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
>>>> You may or may not see output from other processes, depending on
>>>> exactly when Open MPI kills them.
>>>> -------------------------------------------------------------------
>>>> ------- Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> Error: Unsupported datatype passed to
>>>> ADIOI_Count_contiguous_blocks, combiner = 17
>>>> -------------------------------------------------------------------
>>>> ------- mpiexec has exited due to process rank 1 with PID 17756 on
>>>> node XXXX exiting without calling "finalize". This may have caused
>>>> other processes in the application to be terminated by signals sent
>>>> by mpiexec (as reported here)."
>>>>
>>>> As I learnt it has to do something with a particular data type (as
>>>> the error message indicates it), but I'm stuck, I don't know how to
>>>> proceed.
>>>>
>>>> Could You give some help?
>>>>
>>>> Best regards,
>>>> Joseph
>>>>
>>>> _______________________________________________
>>>> Hdf-forum is for HDF software users discussion.
>>>> Hdf-forum@lists.hdfgroup.org
>>>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup
>>>> .org
>>>> Twitter: https://twitter.com/hdf5
>>>
>>> _______________________________________________
>>> Hdf-forum is for HDF software users discussion.
>>> Hdf-forum@lists.hdfgroup.org
>>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup\.
>>> org
>>> Twitter: https://twitter.com/hdf5
>>>
>>
>> -------------------------------------------
>>
>> József Kóbori
>> PhD Student - Particle Physics and Astronomy Program
>> Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös
>> Loránd University
>> 1117 Budapest, Pázmány P. stny. 1/A
>>
>> _______________________________________________
>> Hdf-forum is for HDF software users discussion.
>> Hdf-forum@lists.hdfgroup.org
>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.o
>> rg
>> Twitter: https://twitter.com/hdf5
>
> _______________________________________________
> Hdf-forum is for HDF software users discussion.
> Hdf-forum@lists.hdfgroup.org
> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.or
> g
> Twitter: https://twitter.com/hdf5
>

-------------------------------------------

József Kóbori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös Loránd
University
1117 Budapest, Pázmány P. stny. 1/A

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

-------------------------------------------

József Kóbori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673
Eötvös Loránd University
1117 Budapest, Pázmány P. stny. 1/A

You are using a very old version of Open MPI.
Open MPI: 1.4.5
Open MPI SVN revision: r25905
Open MPI release date: Feb 10, 2012

Please use latest.

Thanks,
Mohamad

···

-----Original Message-----
From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On Behalf
Of Kóbori József
Sent: Wednesday, November 25, 2015 3:46 AM
To: hdf-forum@lists.hdfgroup.org
Subject: Re: [Hdf-forum] unsupported data type

  Hi Mohamad,

please found attached the output of "ompi_info" and the log of the faulty
hdf5 test.
Thanks!

  Regards,
   Joseph

Idézet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):

> Hi Joseph,
>
> What is the test that fails in make check? When you run make check, it
> should tell you which test is running. Please provide that.
> I also would like to know the Open-MPI version you are using.
>
> Thanks,
> Mohamad
>
>> -----Original Message-----
>> From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org] On
>> Behalf Of Kóbori József
>> Sent: Tuesday, November 24, 2015 8:04 AM
>> To: hdf-forum@lists.hdfgroup.org
>> Subject: Re: [Hdf-forum] unsupported data type
>>
>> Hi Richard,
>>
>> I used "configure" with the following options:
>> --prefix=/some/dir --enable-parallel
>>
>> Platform:
>> Linux atlasz 3.2.0-4-amd64 #1 SMP Debian 3.2.68-1+deb7u1 x86_64
>> GNU/Linux
>>
>> The libhdf5.settings file is attached.
>>
>> Thanks!
>>
>> Regards,
>> Joseph
>>
>> Idézet (Richard van Hees <r.m.van.hees@sron.nl>):
>>
>> > Hi Joseph ,
>> >
>> > It would be helpful if you would indicate which way you configured
>> > HDF5. Using cmake? Using configure? Which options did you use? On
>> > which platform?
>> >
>> > Helpful also be the content of the file
>> > <HDF5_DIR>/lib/libhdf5.settings, where <HDF5_DIR> is the
>> > installation directory for HDF5, default "/usr/local" or specified
>> > by you using PREFIX or --prefix
>> >
>> > Greetings, Richard
>> >
>> > On 11/24/2015 12:40 PM, Kóbori József wrote:
>> >> Hi Mohamad,
>> >>
>> >> I compiled the latest version of HDF5: hdf5-1.8.16 After compiling
>> >> (make) I simply typed the "make check" command, and I got the
>> >> error below, so I don't have a particular test case.
>> >> If You need more info please let me know!
>> >> Thank You!
>> >>
>> >> Regards,
>> >> Joseph
>> >>
>> >> Idézet (Mohamad Chaarawi <chaarawi@hdfgroup.org>):
>> >>
>> >>> Hi Joseph,
>> >>>
>> >>> Could you please indicate the version of the HDF5 library and
>> >>> Open-MPI that you are using?
>> >>> Furthermore, could you provide a test case that replicates the
>> >>> failure below?
>> >>>
>> >>> You are correct that it seems that we are constructing a derived
>> >>> datatype that the ADIO driver in ROMIO doesn't seem to be able to
>> >>> parse properly. We don't test however much with Open-MPI, at
>> >>> lease to the extent that we do with MPICH. I'd be curious if you
>> >>> can try the same program with MPICH and see if it gives the same
error.
>> >>>
>> >>> Thanks,
>> >>> Mohamad
>> >>>
>> >>>> -----Original Message-----
>> >>>> From: Hdf-forum [mailto:hdf-forum-bounces@lists.hdfgroup.org]
On
>> >>>> Behalf Of Kóbori József
>> >>>> Sent: Monday, November 16, 2015 5:30 AM
>> >>>> To: hdf-forum@lists.hdfgroup.org
>> >>>> Subject: [Hdf-forum] unsupported data type
>> >>>>
>> >>>> Dear All,
>> >>>>
>> >>>> when trying to test the previously compiled (with mpicc) HDF5
>> >>>> source I encountered the following error message:
>> >>>>
>> >>>> "MPI_ABORT was invoked on rank 3 in communicator
>> MPI_COMM_WORLD
>> >>>> with errorcode 1.
>> >>>>
>> >>>> NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI
processes.
>> >>>> You may or may not see output from other processes, depending
on
>> >>>> exactly when Open MPI kills them.
>> >>>> ----------------------------------------------------------------
>> >>>> ---
>> >>>> ------- Error: Unsupported datatype passed to
>> >>>> ADIOI_Count_contiguous_blocks, combiner = 17
>> >>>> Error: Unsupported datatype passed to
>> >>>> ADIOI_Count_contiguous_blocks, combiner = 17
>> >>>> Error: Unsupported datatype passed to
>> >>>> ADIOI_Count_contiguous_blocks, combiner = 17
>> >>>> Error: Unsupported datatype passed to
>> >>>> ADIOI_Count_contiguous_blocks, combiner = 17
>> >>>> Error: Unsupported datatype passed to
>> >>>> ADIOI_Count_contiguous_blocks, combiner = 17
>> >>>> Error: Unsupported datatype passed to
>> >>>> ADIOI_Count_contiguous_blocks, combiner = 17
>> >>>> ----------------------------------------------------------------
>> >>>> ---
>> >>>> ------- mpiexec has exited due to process rank 1 with PID 17756
>> >>>> on node XXXX exiting without calling "finalize". This may have
>> >>>> caused other processes in the application to be terminated by
>> >>>> signals sent by mpiexec (as reported here)."
>> >>>>
>> >>>> As I learnt it has to do something with a particular data type
>> >>>> (as the error message indicates it), but I'm stuck, I don't know
>> >>>> how to proceed.
>> >>>>
>> >>>> Could You give some help?
>> >>>>
>> >>>> Best regards,
>> >>>> Joseph
>> >>>>
>> >>>>
>> >>>> _______________________________________________
>> >>>> Hdf-forum is for HDF software users discussion.
>> >>>> Hdf-forum@lists.hdfgroup.org
>> >>>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgr
>> >>>> oup
>> >>>> .org
>> >>>> Twitter: https://twitter.com/hdf5
>> >>>
>> >>> _______________________________________________
>> >>> Hdf-forum is for HDF software users discussion.
>> >>> Hdf-forum@lists.hdfgroup.org
>> >>> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.
>> >>> org
>> >>> Twitter: https://twitter.com/hdf5
>> >>>
>> >>
>> >>
>> >>
>> >> -------------------------------------------
>> >>
>> >> József Kóbori
>> >> PhD Student - Particle Physics and Astronomy Program
>> >> Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös
>> >> Loránd University
>> >> 1117 Budapest, Pázmány P. stny. 1/A
>> >>
>> >>
>> >>
>> >> _______________________________________________
>> >> Hdf-forum is for HDF software users discussion.
>> >> Hdf-forum@lists.hdfgroup.org
>> >> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgrou
>> >> p.o
>> >> rg
>> >> Twitter: https://twitter.com/hdf5
>> >
>> >
>> > _______________________________________________
>> > Hdf-forum is for HDF software users discussion.
>> > Hdf-forum@lists.hdfgroup.org
>> > http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup
>> > .or
>> > g
>> > Twitter: https://twitter.com/hdf5
>> >
>>
>>
>>
>> -------------------------------------------
>>
>> József Kóbori
>> PhD Student - Particle Physics and Astronomy Program
>> Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös
>> Loránd University
>> 1117 Budapest, Pázmány P. stny. 1/A
>
> _______________________________________________
> Hdf-forum is for HDF software users discussion.
> Hdf-forum@lists.hdfgroup.org
> http://lists.hdfgroup.org/mailman/listinfo/hdf-forum_lists.hdfgroup.or
> g
> Twitter: https://twitter.com/hdf5
>

-------------------------------------------

József Kóbori
PhD Student - Particle Physics and Astronomy Program
Room: Building North, Number 5.91, Phone: +36 30-266-5673 Eötvös Loránd
University
1117 Budapest, Pázmány P. stny. 1/A