[aspect-devel] Troubles with opening vtk files in Paraview

Fri Jul 28 11:57:47 PDT 2017

Hi Timo,

most importantly, the problem seems to only occur on large number of procs.
It occurs with 256, but not with 64 procs.

1. I only tried the values of 1 and not specifying it at all (I think it
defaulted to 16). Usually it was file 0 that was corrupted, but for 1 of
the models the 0-the file was OK, and 1st was corrupted.
2. If i set it to 1, if works correctly, yes.
3. If I set "number of grouped files" to 0 it works fine too

I'm using openmpi/2.1.1

Best regards,
Lev Karatun.

2017-07-27 14:22 GMT-04:00 Timo Heister <heister at clemson.edu>:

> Lev,
>
> that is a weird situation. To make sure I understand:
> 1. if you set it to a value greater than one, you randomly get invalid
> .vtu files? Is it always file -0000 or are other files affected?
> 2. If you set it to 1, it works correctly?
> 3. Can you check what happens if you set "number of grouped files" to 0?
>
> As this seems to be related to MPI I/O, can you tell us what MPI
> library you are using?
>
>
>
> On Thu, Jul 27, 2017 at 2:08 PM, Lev Karatun <lev.karatun at gmail.com>
> wrote:
> > Hi,
> >
> > after updating Aspect (not sure if it played any role) I only have
> troubles
> > with vtu files if I don't set "number of grouped files" to 1 (so far, at
> > least). The same input files produces NULL symbols in different files
> during
> > different runs.
> > I can, of course, run models with "number of grouped files" set to 1, so
> I
> > don't need immediate help, but I thought it's worth noting (and maybe
> > registering as a bug) anyway.
> >
> > Best regards,
> > Lev Karatun.
> >
> > 2017-07-24 17:08 GMT-04:00 Lev Karatun <lev.karatun at gmail.com>:
> >>
> >> Hi Juliane, Wolfgang,
> >>
> >> it happenned in several different models, but I can't run them 1 more
> time
> >> and compare the output because when I encountered the problem I started
> >> updating Aspect and now I'm having trouble compiling it with Intel
> compilers
> >> (trying to work it out with the guys from cluster support). I can't use
> gcc
> >> because everything on the cluster is compiled with icc, there is not
> even an
> >> MPI wrapper for gcc.
> >> I appreciate your help and I'll let you know once I can check if the
> error
> >> is reproducible.
> >>
> >>
> >>
> >> Best regards,
> >> Lev Karatun.
> >>
> >> 2017-07-21 16:47 GMT-04:00 Juliane Dannberg <dannberg at gfz-potsdam.de>:
> >>>
> >>> Hi Lev,
> >>>
> >>> if I open your file in a browser, I get the same error (which is
> exactly
> >>> where the weird zero characters you mentioned are):
> >>>
> >>> XML Parsing Error: not well-formed
> >>> Location:
> >>> file:///home/dannberg/old_home_arbeit/test/solution-00000.0000.vtu
> >>> Line Number 434, Column 20:    <DataArray type
> >>>
> >>> So it is definitely a problem with the file (not with Paraview). In
> >>> principle, this line should contain the type of the data array
> (something
> >>> like <DataArray type="Float64" Name="viscosity" format="binary">)
> >>>
> >>> Is this a reproducible error, or did it happen only once?
> >>>
> >>> Best,
> >>> Juliane
> >>>
> >>>
> >>> On 07/21/2017 01:43 PM, Lev Karatun wrote:
> >>>
> >>> Hi Wolfgang,
> >>>
> >>> in the file there is a bunch of null symbols, not sure where they came
> >>> from. I definitely didn't hit ctrl+c and Aspect wasn't killed because
> this
> >>> is the output for 0th timestep, there are others. For the same reason,
> I
> >>> don't think I ran out of space. I attached the file just in case.
> >>>
> >>> Best regards,
> >>> Lev Karatun.
> >>>
> >>> 2017-07-21 15:23 GMT-04:00 Wolfgang Bangerth <bangerth at tamu.edu>:
> >>>>
> >>>>
> >>>> Lev,
> >>>>
> >>>>> I'm getting an error when trying to open a solution file in Paraview.
> >>>>> It's strange because I don't even have a C:\DBD folder. I tried new
> and old
> >>>>> versions of Paraview, enabling and disabling "set Number of grouped
> files =
> >>>>> 1", but it didn't help.
> >>>>
> >>>>
> >>>> C:\DBD is the folder in which Paraview was compiled. If you installed
> a
> >>>> binary file you downloaded, then this is not a path on your system.
> >>>>
> >>>>
> >>>>> Did anyone see something like this before?
> >>>>>
> >>>>> ERROR: In
> >>>>> C:\DBD\pvs-x64\paraview\src\paraview\VTK\IO\XMLParser\vtkXMLParser.cxx,
> line
> >>>>> 483
> >>>>>
> >>>>> vtkXMLDataParser (0000000016127230): Error parsing XML in stream at
> >>>>> line 434, column 19, byte index 233472: not well-formed (invalid
> token)
> >>>>
> >>>>
> >>>> It's a pity that paraview doesn't say which file this is in. What
> >>>> typically happens with errors such as this is that the file you're
> trying to
> >>>> read is corrupted. This could be because you hit Ctrl-C while ASPECT
> >>>> happened to be writing that file (or because your cluster's scheduler
> >>>> happened to kill ASPECT just in that moment). It could also be
> because you
> >>>> ran out of file space while ASPECT was running.
> >>>>
> >>>> If you can guess which file is being read right now, take a look at
> line
> >>>> 434 and see how the context around it looks like.
> >>>>
> >>>> Best
> >>>>  W.
> >>>>
> >>>> --
> >>>> ------------------------------------------------------------
> ------------
> >>>> Wolfgang Bangerth          email:
> bangerth at colostate.edu
> >>>>                            www: https://urldefense.proofpoint.
> com/v2/url?u=http-3A__www.math.colostate.edu_-7Ebangerth_&d=DwIFaQ&c=Ngd-
> ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=c08Btfq4m9QEScXN3ZQwLZzzWQE7S8
> CYq1IYuzKV_Zk&m=ZI_xyF0SVGrQ5tAYjg9qFVIEBfzOsFqNOWWY4n7Oj8s&s=
> afuGZlrZZTPbKphYVPYO02pzcVjaWhg4HOUFgjlDNPg&e=
> >>>> _______________________________________________
> >>>> Aspect-devel mailing list
> >>>> Aspect-devel at geodynamics.org
> >>>> https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> geodynamics.org_cgi-2Dbin_mailman_listinfo_aspect-2Ddevel&d=DwIFaQ&c=Ngd-
> ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=c08Btfq4m9QEScXN3ZQwLZzzWQE7S8
> CYq1IYuzKV_Zk&m=ZI_xyF0SVGrQ5tAYjg9qFVIEBfzOsFqNOWWY4n7Oj8s&s=4-
> Gqd9FqQ67Hijizitxbqrk71OazjgvoNAbDeN_CSDw&e=
> >>>
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> Aspect-devel mailing list
> >>> Aspect-devel at geodynamics.org
> >>> https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> geodynamics.org_cgi-2Dbin_mailman_listinfo_aspect-2Ddevel&d=DwIFaQ&c=Ngd-
> ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=c08Btfq4m9QEScXN3ZQwLZzzWQE7S8
> CYq1IYuzKV_Zk&m=ZI_xyF0SVGrQ5tAYjg9qFVIEBfzOsFqNOWWY4n7Oj8s&s=4-
> Gqd9FqQ67Hijizitxbqrk71OazjgvoNAbDeN_CSDw&e=
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> Aspect-devel mailing list
> >>> Aspect-devel at geodynamics.org
> >>> https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> geodynamics.org_cgi-2Dbin_mailman_listinfo_aspect-2Ddevel&d=DwIFaQ&c=Ngd-
> ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=c08Btfq4m9QEScXN3ZQwLZzzWQE7S8
> CYq1IYuzKV_Zk&m=ZI_xyF0SVGrQ5tAYjg9qFVIEBfzOsFqNOWWY4n7Oj8s&s=4-
> Gqd9FqQ67Hijizitxbqrk71OazjgvoNAbDeN_CSDw&e=
> >>
> >>
> >
> >
> > _______________________________________________
> > Aspect-devel mailing list
> > Aspect-devel at geodynamics.org
> > https://urldefense.proofpoint.com/v2/url?u=http-3A__lists.
> geodynamics.org_cgi-2Dbin_mailman_listinfo_aspect-2Ddevel&d=DwIGaQ&c=Ngd-
> ta5yRYsqeUsEDgxhcqsYYY1Xs5ogLxWPA_2Wlc4&r=R5lvg9JC99XvuTgScgbY_
> QFS80R7PEA2q0EPwDy7VQw&m=fHrrR_HHTqGbKx4SlWgFnCbKUYC97CIMrphO
> T3ifKAs&s=qV5f-VHy1WEbJTSSOIOvqjTm8ESoZKreOnsIVes43jc&e=
>
>
>
> --
> Timo Heister
> http://www.math.clemson.edu/~heister/
> _______________________________________________
> Aspect-devel mailing list
> Aspect-devel at geodynamics.org
> http://lists.geodynamics.org/cgi-bin/mailman/listinfo/aspect-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20170728/44020eac/attachment-0001.html>