Thanks Joachim. It was indeed an issue of not using the right mpiexec. On Mon, Oct 16, 2017 at 10:46 AM, Joachim Pouderoux < [email protected]> wrote:
> To be more specific, you have to run something like: > $ /pathToPV5.0/lib/paraview-5.0/mpiexec -np 32 /pathToPV5.0/bin/pvserver > > Joachim > > *Joachim Pouderoux*, PhD > > *Technical Expert - Scientific Computing Team* > *Kitware SAS <http://www.kitware.fr>* > > > 2017-10-16 13:27 GMT-04:00 Joachim Pouderoux < > [email protected]>: > >> Again Quentin, which "mpiexec" program are you using? >> The parallel PVServer is supposed to open just one socket to accept >> client's connection. Otherwise it just means that the MPI context is not >> correctly setup. >> It occures if PVServer has not been built with MPI support OR you do not >> use the mpiexec that correspond to the MPI library used to compile PV. >> >> Joachim >> >> *Joachim Pouderoux*, PhD >> >> *Technical Expert - Scientific Computing Team* >> *Kitware SAS <http://www.kitware.fr>* >> >> >> 2017-10-16 13:23 GMT-04:00 Quentin d'Avout <[email protected]>: >> >>> Hi Joachim, >>> >>> >>> >>> I’m running: >>> >>> mpirun -np 32 pvserver --server-port=%PV_SERVER_PORT% >>> >>> per paraView’s website. Launching the pvserver is not the issue, >>> connecting to it in the paraview GUI is. >>> >>> Thanks. >>> >>> >>> >>> >>> >>> *From:* Joachim Pouderoux [mailto:[email protected]] >>> *Sent:* Monday, October 16, 2017 10:08 AM >>> *To:* Quentin d'Avout <[email protected]> >>> *Cc:* Utkarsh Ayachit <[email protected]>; ParaView < >>> [email protected]> >>> >>> *Subject:* Re: [Paraview] [EXTERNAL] pvserver MPI issue >>> >>> >>> >>> Quentin, >>> >>> Which `mpiexec` command are you using? >>> >>> You have to use the one provided with paraview, not the one of your >>> system which might be incompatible. >>> >>> Best, >>> >>> Joachim >>> >>> >>> *Joachim Pouderoux*, PhD >>> >>> *Technical Expert - Scientific Computing Team* >>> *Kitware SAS <http://www.kitware.fr>* >>> >>> >>> >>> 2017-10-16 13:03 GMT-04:00 Quentin d'Avout <[email protected]>: >>> >>> Hi Utkarsh, >>> >>> I did not recompile paraView myself, and using version >>> ParaView-5.0.1-Qt4-OpenGL2-MPI-Linux-64bit >>> >>> I assumed that because the D3 fitler is available and that the about >>> menus is stating "MPI enabled", this version was MPI enabled. Am I wrong >>> tho? >>> >>> [image: Inline image 1] >>> >>> Thanks >>> >>> Quentin >>> >>> >>> >>> On Sat, Oct 14, 2017 at 5:23 PM, Utkarsh Ayachit < >>> [email protected]> wrote: >>> >>> Quentin, >>> >>> >>> >>> Each pvserver rank trying to connect the same socket is the classic >>> symptom of paraview not build with MPI. Mind attaching your CMakeCache.txt? >>> >>> >>> >>> Utkarsh >>> >>> >>> >>> On Sat, Oct 14, 2017 at 8:05 PM, Quentin d'Avout < >>> [email protected]> wrote: >>> >>> Thanks Utkarsh. >>> >>> Yes, I’ve checked both. >>> >>> >>> >>> Quentin d’Avout | Aerodynamics >>> >>> [image: Arrivo Loop] <http://www.arrivo-loop.com/> >>> >>> >>> >>> *From:* Utkarsh Ayachit [mailto:[email protected]] >>> *Sent:* Saturday, October 14, 2017 4:55 PM >>> *To:* Quentin d'Avout <[email protected]> >>> *Cc:* Scott, W Alan <[email protected]>; ParaView < >>> [email protected]> >>> *Subject:* Re: [Paraview] [EXTERNAL] pvserver MPI issue >>> >>> >>> >>> Are you sure you ParaView is built with MPI support enabled? If so, are >>> you sure you're using the correct mpirun executable that goes with the MPI >>> implementation you used to build ParaView? >>> >>> >>> >>> Utkarsh >>> >>> >>> >>> On Sat, Oct 14, 2017 at 2:49 PM, Quentin d'Avout < >>> [email protected]> wrote: >>> >>> Actually, it turns out that you have to specifically tell pvserver to >>> use different ports for every process: >>> >>> >>> >>> mpirun -np 32 pvserver --server-port=%PV_SERVER_PORT% >>> >>> >>> >>> But I’m now having issues connecting to all of these processes from the >>> paraView GUI/connect window. Can’t seem to be able to specify more than one >>> port at once… >>> >>> >>> >>> Anybody knows the trick? >>> >>> Thanks again >>> >>> >>> >>> *From:* Scott, W Alan [mailto:[email protected]] >>> *Sent:* Friday, October 13, 2017 6:40 PM >>> >>> >>> *To:* Quentin d'Avout <[email protected]>; [email protected] >>> *Subject:* Re: [EXTERNAL] [Paraview] pvserver MPI issue >>> >>> >>> >>> All I have is back to the firewall? >>> >>> >>> >>> *From: *Quentin d'Avout <[email protected]> >>> *Date: *Friday, October 13, 2017 at 4:39 PM >>> *To: *W Scott <[email protected]>, "[email protected]" < >>> [email protected]> >>> *Subject: *RE: [EXTERNAL] [Paraview] pvserver MPI issue >>> >>> >>> >>> Thanks Scott. >>> >>> Didn’t do the trick tho. >>> >>> Other ideas? >>> >>> >>> >>> *From:* Scott, W Alan [mailto:[email protected]] >>> *Sent:* Friday, October 13, 2017 2:15 PM >>> *To:* Quentin d'Avout <[email protected]>; [email protected] >>> *Subject:* Re: [EXTERNAL] [Paraview] pvserver MPI issue >>> >>> >>> >>> Not saying I’m right here, but it looks like multiple versions of >>> ParaView are tying up the socket. Do a ps -ef on paraview, client side, >>> and kill them. Do the same on the server side. Then, try again. >>> >>> >>> >>> Alan >>> >>> >>> >>> *From: *ParaView <[email protected]> on behalf of Quentin >>> d'Avout <[email protected]> >>> *Date: *Friday, October 13, 2017 at 3:10 PM >>> *To: *"[email protected]" <[email protected]> >>> *Subject: *[EXTERNAL] [Paraview] pvserver MPI issue >>> >>> >>> >>> Hi All, >>> >>> >>> >>> my issue is when I hit: >>> >>> >>> >>> mpirun -np 32 pvserver # with or without --mpi option >>> >>> >>> >>> I get :"vtkServerSocket (0x23fb3e0): Socket error in call to bind. >>> Address already in use." >>> >>> and: "vtkTCPNetworkAccessManager (0x1879510): Failed to set up server >>> socket." >>> >>> >>> >>> as if my paraView was compiled with the mpi capability off. But it is on >>> (D3 filter available, "about" windows also says so). >>> >>> >>> >>> pvserver in serial works just fine. >>> >>> Could have been the ufw firewall blocking connections, but it is >>> disabled. >>> >>> Ubuntu 17.04. >>> >>> >>> >>> Thanks in advance for your help. >>> >>> Quentin >>> >>> >>> _______________________________________________ >>> 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 >>> >>> >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> 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 >>> >>> >>> >> >> > -- Quentin d’Avout | Aerodynamics [image: Arrivo Loop] <http://www.arrivo-loop.com/>
_______________________________________________ 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
