[aspect-devel] problem on output writing

FELIPE ORELLANA f_orellana at berkeley.edu
Mon Feb 8 12:09:00 PST 2016


   Dear Aspect community,

        I am running ASPECT using 64 nodes on a remote cluster.

        I keep getting a problem with output writing. It happens that after
some time steps my system, although it runs without complaining about
failure, gives a warning and it does not continue writing output vtu, pvtu
files

 this is an example message i get on a log file:

  Solving temperature system... 13 iterations.
   Solving Stokes system... ***** WARNING: could not move
/tmp/aspect.tmp.1L4hx2 to /home/felipe/ouRB/solution-00000.0019.vtu. Trying
again to write directly to /home/felipe/ouRB/solution-00000.0019.vtu. (
On processor 19.)
***** WARNING: could not move /tmp/aspect.tmp.Tb62qG to
/home/felipe/ouRB/solution-00000.0044.vtu. Trying again to write directly
to /home/felipe/ouRB/solution-00000.0044.vtu. (On processor 44.)
***** WARNING: could not move /tmp/aspect.tmp.vkgo6H to
/home/felipe/ouRB/solution-00000.0033.vtu. Trying again to write directly
to /home/felipe/ouRB/solution-00000.0033.vtu. (On processor 33.)
***** WARNING: could not move /tmp/aspect.tmp.H0pvlY to
/home/felipe/ouRB/solution-00000.0049.vtu. Trying again to write directly
to /home/felipe/ouRB/solution-00000.0049.vtu. (On processor 49.)
***** WARNING: could not move /tmp/aspect.tmp.vcfWPj to
/home/felipe/ouRB/solution-00000.0013.vtu. Trying again to write directly
to /home/felipe/ouRB/solution-00000.0013.vtu. (On processor 13.)

 I wonder if this is related to what is written on section 4.4.3 of
ASPECT's handbook about the cluster's capabilities. For example, here there
is a subsection of it:

  ''If your cluster has a slow interconnect, e.g., if it is simply a
collection of machines connected via Ethernet, then writing data to a
central file server may block the rest of the program for a while. On the
other hand, if your machines have fast local storage for temporary file
systems, then ASPECT can write data first into such a file and then move it
in the background to its final destination while already continuing
computations. To select this mode, set the appropriate variables discussed
in Section 5.74. Note, however, that this scheme only makes sense if every
machine on which MPI processes run has fast local disk space for temporary
storage.''

what should I do?
I have tried several things unsuccessfully.

I appreciate your comprehension,
cheers,

Felipe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geodynamics.org/pipermail/aspect-devel/attachments/20160208/0d47525d/attachment.html>


More information about the Aspect-devel mailing list