[mpich-discuss] Problems using MPI_File functions

[Pablo Guaza Peces]

Hmmm,
How could I do that?!
The thing is that I'm not really used to program in this kind of 
environment, I'm the person who takes care of the system installation 
and set up for the users, but doesn't use it...

:-/

thanks!

El mié 22 ago 2012 16:57:45 CEST, Rob Latham escribió:
> On Wed, Aug 22, 2012 at 01:48:29PM +0200, Pablo Guaza Peces wrote:
>> Hi everybody!
>> I'm having some problems using MPI_File functions. I made a very
>> simple program to test it after one of my users had some issues with
>> a more complicated program.
>>
>> This is the program I'm using:
>
> sure seems like a simple enough program.
>
>> If I run this program with a PBS script only using one of my nodes
>> it works well, but if I use two nodes, it creates the file but
>> afterwards it just keeps running, and I have to kill it.
>>
>> I think this might has something to do with the NFS share that the
>> programs is using to create the file, but I don't really know. Could
>> it be a NFS locking issue?
>
> It's true we don't pay a lot of attention to NFS, but I can't
> reproduce the problem here.  Any chance you can attach a debugger to
> the stuck processor and let me know what the backtrace looks like?
>
> thanks
> ==rob
>