[Paraview] PV 5.0.1 client-server crashing for some datasets

[Matthew Brown]

I rebuilt the cluster's installation of Paraview 5.0.1 making sure to set
VTK_RENDERING_BACKEND=OpenGL2 since this will be most compatible with the
superbuild binaries.  The datasets I originally could not load now seem to
load and render correctly.  I will have more users test the installation
and report back here if I hear of any problems.

Thanks again,

Matt

On Wed, May 25, 2016 at 3:45 PM, Matthew Brown <brownm12 at vt.edu> wrote:

> Thank you for the quick reply.
>
> On the client, I used a superbuild for MacOS X which I take it is built
> with OpenGL2.  On the cluster, I accepted the default flag which (upon
> rerunning ccmake) appears to be OpenGL on this system.
>
> So, the client and server are almost certainly built to use different
> rendering backends.  I take it that they must be the same for full
> functionality, is that correct?  What functionality would you expect to
> fail if rendering backends are mismatched between the client and server?
>
> Matt
>
> On Wed, May 25, 2016 at 2:55 PM, Utkarsh Ayachit <
> utkarsh.ayachit at kitware.com> wrote:
>
>> Matt,
>>
>> Can you confirm that both the client and server are using the same
>> rendering backend? Unfortunately, there's no clean way to do that
>> except to look at the CMakeCache.txt used to build to two. Confirm
>> that the VTK_RENDERING_BACKEND CMake flag is set to the same value on
>> both. I suspect one is "OpenGL" and other is "OpenGL2".
>>
>> Utkarsh
>>
>> On Wed, May 25, 2016 at 2:04 PM, Matthew Brown <brownm12 at vt.edu> wrote:
>> > I have recently built and installed ParaView 5.0.1 on one of our
>> clusters
>> > and am having issues when loading most datasets that I have tried.
>> >
>> > 1. remote rendering seems to work with no issues:  I can make spheres
>> and
>> > cylinders and move them around and color their sections according the
>> core
>> > doing the processing
>> > 2. the ParaView tutorial dataset (from
>> > http://www.paraview.org/Wiki/The_ParaView_Tutorial) named headsq.vti
>> can
>> > load and render remotely with no problems
>> > 3. loading either of the other two tutorial datasets (can.ex2 or
>> > disk_out_ref.ex2) causes the application to crash
>> > 4. The remote client machine (running ParaView5.0.1) can open all of
>> these
>> > datasets with no problem
>> >
>> > Crash details:
>> > a. data set variables show up correctly in the Properties pane, but
>> > Information seems unavailable: all the fields under statistics and data
>> > arrays have values "NA", but time indices look correct
>> > b. clicking "Apply" causes the client application on the remote machine
>> and
>> > pvserver on the cluster to die
>> > c. client-side standard output:
>> > ####################-begin-####################
>> > $ ./paraview
>> > AutoMPI: SUCCESS: command is:
>> >  "/Applications/paraview.app/Contents/MacOS/mpiexec" "-np" "2"
>> > "/Applications/paraview.app/Contents/bin/pvserver" "--server-port=56934"
>> > AutoMPI: starting process server
>> > -------------- server output --------------
>> > Waiting for client...
>> > AutoMPI: server successfully started.
>> > ERROR: In
>> >
>> /Users/kitware/buildbot-slave/paraview-bigmac-osx-shared-release_opengl2_osx10_7_qt4_superbuild/build/paraview/src/paraview/ParaViewCore/ServerImplementation/Core/vtkSIProxy.cxx,
>> > line 310
>> > vtkSIProxy (0x7f93f3a2ead0): Failed to create vtkHAVSVolumeMapper.
>> Aborting
>> > for debugging purposes.
>> >
>> >
>> > Abort trap: 6
>> > ####################-end-####################
>> >
>> > d. cluster-side standard output:
>> > ####################-begin-####################
>> > [brownm12 at nr019 ~]$ mpirun -n 2 pvserver
>> > Waiting for client...
>> > Connection URL: cs://nr019:11111
>> > Accepting connection(s): nr019:11111
>> > Client connected.
>> > ERROR: In
>> >
>> /opt/build/BUILD/ParaView-5.0.1-Source/VTK/Parallel/Core/vtkSocketCommunicator.cxx,
>> > line 809
>> > vtkSocketCommunicator (0x2f81200): Could not receive tag. 1
>> >
>> > Exiting...
>> > Exiting...
>> > ####################-end-####################
>> >
>> > Build information:
>> > cluster: CentOS 7, gcc-4.7.2, openmpi-1.8.5, python 2.7.10, ParaView
>> 5.0.1.
>> > I can provide more details if needed, but I'm not sure what would be
>> most
>> > relevant
>> >
>> > It seems to me that the issue may be with how the cluster is opening the
>> > datasets.  I would like some advice or ideas on how to resolve or
>> continue
>> > troubleshooting the problem.
>> >
>> > Thanks,
>> >
>> > Matt
>> >
>> >
>> >
>> > _______________________________________________
>> > 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
>> >
>> > Search the list archives at: http://markmail.org/search/?q=ParaView
>> >
>> > Follow this link to subscribe/unsubscribe:
>> > http://public.kitware.com/mailman/listinfo/paraview
>> >
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/paraview/attachments/20160525/f1fe156b/attachment.html>