[MPICH] Behavour if MPI_File_Open fails on some nodes

Robert Latham robl at mcs.anl.gov
Mon Aug 20 16:53:41 CDT 2007

On Mon, Aug 20, 2007 at 02:57:55PM +0100, James S Perrin wrote:
> Case 2 would be fine, however MPI_File_Open (rank 0) is not
> returning at all in the situation I've described.

That's odd.  I ran your test program, creating a file on one processor
that did not exist on others, and got the expcected behavior
(MPI_File_open returns an error on all processes. no process hangs).
Wonder what's different about our environments?   
> BTW since I don't know if the file is shared or duplicated I'm only
> using MPI_File_Open, MPI_File_view, MPI_File_read and
> MPI_File_close. ie nothing that should depended on the shared file
> pointer.

In this context, "shared file pointer" refers to how the file pointer
offset is incremented when a file is accessed.  It doesn't speak to
the availabilty of the file on all processes.   


Rob Latham
Mathematics and Computer Science Division    A215 0178 EA2D B059 8CDF
Argonne National Lab, IL USA                 B29D F333 664A 4280 315B

More information about the mpich-discuss mailing list