[Paraview] ParaView 4.2.0 client/server execution error
Praveen Narayanan
praveenn at nvidia.com
Sat Oct 11 22:30:30 EDT 2014
I haven't worked on client-server mode, but building Paraview with
OSMesa was a very challenging exercise, and I used it for my
workstation. Not sure if these notes would help here. Nevertheless, it
may be pertinent for installation notes.
1) Build OSMesa (obtained from the paraview wiki, and this is for
vanilla OsMesa).
$ ./configure CXXFLAGS=-O2 -g -DDEFAULT_SOFTWARE_DEPTH_BITS=31
CFLAGS=-O2 -g -DDEFAULT_SOFTWARE_DEPTH_BITS=31 --disable-xvmc
--disable-glx --disable-dri --with-dri-drivers= --with-gallium-drivers=
--enable-texture-float --disable-shared-glapi --disable-egl
--with-egl-platforms= --enable-osmesa --enable-gallium-llvm=no
--prefix=/home/praveen/software/osmesa_vanilla_install
2) Use ccmake+superbuild to build paraview, configuring repeatedly. I
had managed several paraview builds with my system mpi (from ubuntu) but
somehow, on that particular day, cmake rejected all of them, and hence
the superbuild install asking it to download mpi and build.
3) For some reason, it would default to using 2 MPI tasks, which we
should change in the ncurses menu. Likewise, the max number of threads
is specified as 64. I am not sure what this would imply.
4) I think it is enough to turn on the OPENGL_HAS_OSMESA flag for it to
use OsMesa.
5) Next, we specify the path for the OsMesa installation (we can point
to any of the OsMesa builds, such as Gallium or Nouveau).
If I recall correctly, ccmake complains if we turn on the GUI. I wonder
if these failures have something to do with the flavor of OsMesa used.
On 10/10/2014 07:37 PM, paraview-request at paraview.org wrote:
>
> Message: 2
> Date: Fri, 10 Oct 2014 22:09:42 +0000
> From: "Biddiscombe, John A." <biddisco at cscs.ch>
> To: "Su, Simon M CTR USARMY ARL (US)" <simon.m.su.ctr at mail.mil>,
> "paraview at paraview.org" <paraview at paraview.org>
> Subject: Re: [Paraview] ParaView 4.2.0 client/server execution error
> on SGI UV System (UNCLASSIFIED)
> Message-ID: <D05E2543.6AB21%biddisco at cscs.ch>
> Content-Type: text/plain; charset="utf-8"
>
> The client has created the representation glyph3d, but the server hasn?t
> got it compiled in by the looks of it, so the create representation fails.
> Does it get skipped when using osmesa (and hence doesn?t exist on the
> server)?
>
> I?m not certain of this, but it happens all the time with the same error
> when one of my plugins that create a representation is loaded on the
> client - but not on the server.
>
> JB
>
> On 10/10/14 20:31, "Su, Simon M CTR USARMY ARL (US)"
> <simon.m.su.ctr at mail.mil> wrote:
>
>> Classification: UNCLASSIFIED
>> Caveats: NONE
>>
>> Hello,
>>
>> We are wondering if someone can point us in the right direction to track
>> down the issue we are getting. We managed to get ParaView OSMESA compile
>> on
>> our SGI UV System. But in a client/server execution mode, although we are
>> able to connect and paraview starts up on the HPC side, as soon as we try
>> to
>> draw something, we get a segmentation fault error - please see error
>> message
>> at the bottom of this email.
>>
>> ParaView 4.2.0 osmesa is configure with the following options.
>>
>> -DCMAKE_BUILD_TYPE=Release \
>> -DParaView_FROM_GIT=ON \
>> -DENABLE_boost=ON \
>> -DENABLE_cgns=OFF \
>> -DENABLE_ffmpeg=ON \
>> -DENABLE_fontconfig=ON \
>> -DENABLE_freetype=ON \
>> -DENABLE_hdf5=ON \
>> -DENABLE_libxml2=ON \
>> -DENABLE_matplotlib=ON \
>> -DENABLE_mesa=OFF \
>> -DENABLE_mpi=ON \
>> -DENABLE_numpy=ON \
>> -DENABLE_osmesa=ON \
>> -DENABLE_paraview=ON \
>> -DENABLE_png=ON \
>> -DENABLE_python=ON \
>> -DENABLE_qhull=ON \
>> -DENABLE_qt=OFF \
>> -DENABLE_silo=ON \
>> -DENABLE_szip=ON \
>> -DENABLE_visitbridge=ON \
>> -DMPI_CXX_LIBRARIES:STRING="/opt/sgi/mpt/mpt-2.08/lib/libmpi++.so" \
>> -DMPI_C_LIBRARIES:STRING="/opt/sgi/mpt/mpt-2.08/lib/libmpi.so" \
>> -DMPI_LIBRARY:FILEPATH="/opt/sgi/mpt/mpt-2.08/lib/libmpi++.so" \
>> -DMPI_CXX_INCLUDE_PATH:STRING="/opt/sgi/mpt/mpt-2.08/include" \
>> -DMPI_C_INCLUDE_PATH:STRING="/opt/sgi/mpt/mpt-2.08/include" \
>> -DUSE_SYSTEM_mpi=ON \
>> -DUSE_SYSTEM_python=OFF \
>> -DUSE_SYSTEM_zlib=OFF \
>>
>> And pvserver is invoked with the following arguments
>>
>> mpiexec_mpt -np ${NUM_MPITASKS} pvserver --use-offscreen-rendering
>> --server-port=1486 --client-host=${HOSTNAME} --reverse-connection
>> --timeout=00:10:00 --connect-id=43826
>>
>> Any suggestion on what/where we can start to track this segmentation fault
>> issue down is much appreciated.
>>
>> Thanks
>> -simon
>>
>>
>> =========================== client/server execution error
>> ========================
>>
>> mpiexec_mpt -np ${NUM_MPITASKS} pvserver --use-offscreen-rendering
>> --server-port=1486 --client-host=predator --reverse-connection
>> --timeout=00:10:00 --connect-id=43826
>>
>> Accepting connection(s): phineas.arl.army.mil:55816
>> ERROR: In
>> /Users/kitware/Dashboards/MyTests/NightlyMaster/ParaViewSuperbuild-Release
>> -
>> Python27/paraview/src/paraview/ParaViewCore/ServerImplementation/Core/vtkP
>> V
>> SessionCore.cxx, line 390
>> vtkPVSessionCore (0x7f8f85b63810): Object type: vtkGlyph3DRepresentation,
>> could not find requested method: "SetTriangulate"
>> or the method was called with incorrect arguments.
>>
>> while processing
>> Message 0 = Invoke
>> Argument 0 = vtk_object_pointer {vtkGlyph3DRepresentation
>> (0x7f8f8833fe90)}
>> Argument 1 = string_value {SetTriangulate}
>> Argument 2 = int32_value {0}
>>
>>
>>
>> ERROR: In
>> /Users/kitware/Dashboards/MyTests/NightlyMaster/ParaViewSuperbuild-Release
>> -
>> Python27/paraview/src/paraview/ParaViewCore/ServerImplementation/Core/vtkP
>> V
>> SessionCore.cxx, line 391
>> vtkPVSessionCore (0x7f8f85b63810): Aborting execution for debugging
>> purposes.
>>
>>
>> ############ ABORT #############
>> ERROR: In
>> /Users/kitware/Dashboards/MyTests/NightlyMaster/ParaViewSuperbuild-Release
>> -
>> Python27/paraview/src/paraview/ParaViewCore/ServerImplementation/Core/vtkS
>> I
>> Proxy.cxx, line 487
>> vtkSISourceProxy (0x7f8f8833fc10): Could not parse property: Triangulate
>>
>>
>> ERROR: In
>> /Users/kitware/Dashboards/MyTests/NightlyMaster/ParaViewSuperbuild-Release
>> -
>> Python27/paraview/src/paraview/ParaViewCore/ClientServerCore/Core/vtkPVAlg
>> o
>> rithmPortsInformation.cxx, line 56
>> vtkPVAlgorithmPortsInformation (0x7f8f88367ee0): Could not downcast
>> vtkAlgorithm.
>>
>>
>> ERROR: In
>> /Users/kitware/Dashboards/MyTests/NightlyMaster/ParaViewSuperbuild-Release
>> -
>> Python27/paraview/src/paraview/ParaViewCore/ServerImplementation/Core/vtkP
>> V
>> SessionCore.cxx, line 390
>> vtkPVSessionCore (0x7f8f85b63810): Wrapper function not found for class
>> "(vtk object is NULL)".
>> while processing
>> Message 0 = Invoke
>> Argument 0 = stream_value {
>> Message 0 = Invoke
>> Argument 0 = id_value {1}
>> Argument 1 = string_value {GetVTKObject}
>> Argument 2 = uint32_value {4525}
>> }
>> Argument 1 = string_value {Initialize}
>> Argument 2 = uint32_value {4525}
>> Argument 3 = uint32_value {4535}
>>
>>
>>
>> ERROR: In
>> /Users/kitware/Dashboards/MyTests/NightlyMaster/ParaViewSuperbuild-Release
>> -
>> Python27/paraview/src/paraview/ParaViewCore/ServerImplementation/Core/vtkP
>> V
>> SessionCore.cxx, line 391
>> vtkPVSessionCore (0x7f8f85b63810): Aborting execution for debugging
>> purposes.
>>
>>
>> ############ ABORT #############
>> Segmentation fault
>>
>>
>>
>> Classification: UNCLASSIFIED
>> Caveats: NONE
>>
>>
>
> ------------------------------
>
> Message: 3
> Date: Fri, 10 Oct 2014 21:07:41 -0400
> From: Cory Quammen <cory.quammen at kitware.com>
> To: "B.W.H. van Beest" <bwvb at xs4all.nl>
> Cc: ParaView <paraview at paraview.org>
> Subject: Re: [Paraview] Bug in Paraview 4.2 example + fix
> Message-ID:
> <CAB5Fpx4gxQGk0bDuYugD2EyiJ6T35g82ua0z=dp8u3TN_6T4pA at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
>
> Hi Bertwin,
>
> No problem. It would be helpful to me, though, if you could checkout
> the change, try it out, and report back on the list if it works for
> you.
>
> To get the change, you will need to have obtained the source via git.
> If this is the case, go to your ParaView source directory and enter
>
> git fetch http://review.source.kitware.com/p/ParaView refs/topics/37/4837/1
> git checkout FETCH_HEAD -b 15044_update_example_plugins
>
> This will download my topic branch and check it out in your local
> repository. Compile it again and test.
>
> Thanks,
> Cory
>
> On Fri, Oct 10, 2014 at 3:36 PM, B.W.H. van Beest <bwvb at xs4all.nl> wrote:
>> Hi Cory,
>>
>> Thanks for putting the fix in. Though willing to contribute, I have no
>> experience with gerrit. So I don't know what to do, really.
>>
>> Regards,
>> Bertwim
>>
>>
>> On 10/10/2014 08:47 PM, Cory Quammen wrote:
>>> Hi Bertwim,
>>>
>>> Thanks for reporting this bug and proposing a fix.
>>>
>>> I posted your report as a bug in the bug tracker:
>>>
>>> http://www.paraview.org/Bug/view.php?id=15044
>>>
>>> I also pushed a git topic to gerrit with your fix and some fixes for
>>> the HiddenLinesRemoval plugin. Please feel free to review it:
>>>
>>> http://review.source.kitware.com/#/t/4837/
>>>
>>> Thanks,
>>> Cory
>>>
>>> On Fri, Oct 10, 2014 at 1:53 PM, B.W.H. van Beest <bwvb at xs4all.nl> wrote:
>>>> Could perhaps somebody of Kitware react to this posting of a bug?
>>>> I noticed a similar issue with the Example plugin "HiddenLinesRemoval",
>>>> although in that case there seems to be more wrong.
>>>>
>>>> Regards,
>>>> Bertwim
>>>>
>>>>
>>>> On 10/07/2014 10:30 PM, B.W.H. van Beest wrote:
>>>>> Hello,
>>>>>
>>>>> I think there is a bug in the plugin example in
>>>>> ../ParaView/Examples/Plugin/Representation (Paraview 4.2)
>>>>>
>>>>> As given, loading the plugin manually, and then creating a new source
>>>>> (e.g. Box) causes Paraview to crash with the error message:
>>>>>
>>>>>> ERROR: In
>>>>>> /usr/local/git/ParaView/ParaViewCore/ServerImplementation/Core/vtkSIProxyDefinitionManager.cxx,
>>>>>> line 565
>>>>>> vtkSIProxyDefinitionManager (0x265ebb0): No proxy that matches:
>>>>>> group=representations and proxy=SurfaceRepresentationBase were found.
>>>>>>
>>>>>>
>>>>>> ERROR: In
>>>>>> /usr/local/git/ParaView/ParaViewCore/ServerImplementation/Core/vtkSIProxyDefinitionManager.cxx,
>>>>>> line 989
>>>>>> vtkSIProxyDefinitionManager (0x265ebb0): Failed to locate base proxy
>>>>>> definition (representations, SurfaceRepresentationBase). Aborting for
>>>>>> debugging purposes.
>>>>>>
>>>>>>
>>>>>> Abort
>>>>> The error seems to be in the file Representation.xml, where the attribute
>>>>> base_proxyname is set to "SurfaceRepresentationBase", which should be
>>>>> "SurfaceRepresentation" according to the documentation in
>>>>> http://www.paraview.org/Wiki/ParaView/Plugin_HowTo#Adding_new_Representations_for_3D_View_using_Plugins_.2A_new_in_version_3.7
>>>>>
>>>>> When I make the change, recompile the plugin and load in into Paraview,
>>>>> there is no longer the above crash when I create a new Box, and the
>>>>> representation "Special Mapper" feautures in the Representationnn combobox.
>>>>>
>>>>>> [243] Representation> diff Representation.xml.org Representation.xml
>>>>>> 7c7
>>>>>> < base_proxyname="SurfaceRepresentationBase">
>>>>>> ---
>>>>>>> base_proxyname="SurfaceRepresentation">
>>>>> Kind Regards,
>>>>> Bertwim
>>>>>
>>>>> _______________________________________________
>>>>> Powered by www.kitware.com
>>>>>
>>>>> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>>>>>
>>>>> Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://public.kitware.com/mailman/listinfo/paraview
>>>>>
>>>>>
>>>> _______________________________________________
>>>> Powered by www.kitware.com
>>>>
>>>> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>>>>
>>>> Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>>>
>>>> Follow this link to subscribe/unsubscribe:
>>>> http://public.kitware.com/mailman/listinfo/paraview
>
> ------------------------------
>
> Message: 4
> Date: Sat, 11 Oct 2014 04:37:13 +0200
> From: "B.W.H. van Beest" <bwvb at xs4all.nl>
> To: Cory Quammen <cory.quammen at kitware.com>
> Cc: ParaView <paraview at paraview.org>
> Subject: Re: [Paraview] Bug in Paraview 4.2 example + fix
> Message-ID: <543897D9.2040102 at xs4all.nl>
> Content-Type: text/plain; charset=utf-8
>
> Cory:
>
> I did as you explained. The fixes work:
> After 'make clean' in the examples directory and 'make all';
>
> Tools -> Manage Plugins -> load libRepresentation.so
> 1. The representation plugin works correctly (After loading e.g. Box
> source, gives the representation "Special Mapper")
>
> Tools -> Manage Plugins -> load libHiddenLinesRemoval.so
> 2. The HiddenLinesRemoval plugin works correctly as well. (After loading
> e.g. Box source, gives the representation "Visible Wireframe", which
> works as expected)
>
> Regards,
> Bertwim
>
>
> On 10/11/2014 03:07 AM, Cory Quammen wrote:
>> Hi Bertwin,
>>
>> No problem. It would be helpful to me, though, if you could checkout
>> the change, try it out, and report back on the list if it works for
>> you.
>>
>> To get the change, you will need to have obtained the source via git.
>> If this is the case, go to your ParaView source directory and enter
>>
>> git fetch http://review.source.kitware.com/p/ParaView refs/topics/37/4837/1
>> git checkout FETCH_HEAD -b 15044_update_example_plugins
>>
>> This will download my topic branch and check it out in your local
>> repository. Compile it again and test.
>>
>> Thanks,
>> Cory
>>
>> On Fri, Oct 10, 2014 at 3:36 PM, B.W.H. van Beest <bwvb at xs4all.nl> wrote:
>>> Hi Cory,
>>>
>>> Thanks for putting the fix in. Though willing to contribute, I have no
>>> experience with gerrit. So I don't know what to do, really.
>>>
>>> Regards,
>>> Bertwim
>>>
>>>
>>> On 10/10/2014 08:47 PM, Cory Quammen wrote:
>>>> Hi Bertwim,
>>>>
>>>> Thanks for reporting this bug and proposing a fix.
>>>>
>>>> I posted your report as a bug in the bug tracker:
>>>>
>>>> http://www.paraview.org/Bug/view.php?id=15044
>>>>
>>>> I also pushed a git topic to gerrit with your fix and some fixes for
>>>> the HiddenLinesRemoval plugin. Please feel free to review it:
>>>>
>>>> http://review.source.kitware.com/#/t/4837/
>>>>
>>>> Thanks,
>>>> Cory
>>>>
>>>> On Fri, Oct 10, 2014 at 1:53 PM, B.W.H. van Beest <bwvb at xs4all.nl> wrote:
>>>>> Could perhaps somebody of Kitware react to this posting of a bug?
>>>>> I noticed a similar issue with the Example plugin "HiddenLinesRemoval",
>>>>> although in that case there seems to be more wrong.
>>>>>
>>>>> Regards,
>>>>> Bertwim
>>>>>
>>>>>
>>>>> On 10/07/2014 10:30 PM, B.W.H. van Beest wrote:
>>>>>> Hello,
>>>>>>
>>>>>> I think there is a bug in the plugin example in
>>>>>> ../ParaView/Examples/Plugin/Representation (Paraview 4.2)
>>>>>>
>>>>>> As given, loading the plugin manually, and then creating a new source
>>>>>> (e.g. Box) causes Paraview to crash with the error message:
>>>>>>
>>>>>>> ERROR: In
>>>>>>> /usr/local/git/ParaView/ParaViewCore/ServerImplementation/Core/vtkSIProxyDefinitionManager.cxx,
>>>>>>> line 565
>>>>>>> vtkSIProxyDefinitionManager (0x265ebb0): No proxy that matches:
>>>>>>> group=representations and proxy=SurfaceRepresentationBase were found.
>>>>>>>
>>>>>>>
>>>>>>> ERROR: In
>>>>>>> /usr/local/git/ParaView/ParaViewCore/ServerImplementation/Core/vtkSIProxyDefinitionManager.cxx,
>>>>>>> line 989
>>>>>>> vtkSIProxyDefinitionManager (0x265ebb0): Failed to locate base proxy
>>>>>>> definition (representations, SurfaceRepresentationBase). Aborting for
>>>>>>> debugging purposes.
>>>>>>>
>>>>>>>
>>>>>>> Abort
>>>>>> The error seems to be in the file Representation.xml, where the attribute
>>>>>> base_proxyname is set to "SurfaceRepresentationBase", which should be
>>>>>> "SurfaceRepresentation" according to the documentation in
>>>>>> http://www.paraview.org/Wiki/ParaView/Plugin_HowTo#Adding_new_Representations_for_3D_View_using_Plugins_.2A_new_in_version_3.7
>>>>>>
>>>>>> When I make the change, recompile the plugin and load in into Paraview,
>>>>>> there is no longer the above crash when I create a new Box, and the
>>>>>> representation "Special Mapper" feautures in the Representationnn combobox.
>>>>>>
>>>>>>> [243] Representation> diff Representation.xml.org Representation.xml
>>>>>>> 7c7
>>>>>>> < base_proxyname="SurfaceRepresentationBase">
>>>>>>> ---
>>>>>>>> base_proxyname="SurfaceRepresentation">
>>>>>> Kind Regards,
>>>>>> Bertwim
>>>>>>
>>>>>> _______________________________________________
>>>>>> Powered by www.kitware.com
>>>>>>
>>>>>> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>>>>>>
>>>>>> Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>>>>>
>>>>>> Follow this link to subscribe/unsubscribe:
>>>>>> http://public.kitware.com/mailman/listinfo/paraview
>>>>>>
>>>>>>
>>>>> _______________________________________________
>>>>> Powered by www.kitware.com
>>>>>
>>>>> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>>>>>
>>>>> Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://public.kitware.com/mailman/listinfo/paraview
>
>
> ------------------------------
>
> Subject: Digest Footer
>
> _______________________________________________
> Powered by www.kitware.com
>
> Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html
>
> Please keep messages on-topic and check the ParaView Wiki at: http://paraview.org/Wiki/ParaView
>
> Follow this link to subscribe/unsubscribe:
> http://public.kitware.com/mailman/listinfo/paraview
>
>
> ------------------------------
>
> End of ParaView Digest, Vol 126, Issue 37
> *****************************************
-----------------------------------------------------------------------------------
This email message is for the sole use of the intended recipient(s) and may contain
confidential information. Any unauthorized review, use, disclosure or distribution
is prohibited. If you are not the intended recipient, please contact the sender by
reply email and destroy all copies of the original message.
-----------------------------------------------------------------------------------
More information about the ParaView
mailing list