[mpich-discuss] Hydra SIGTSTP (Ctrl+Z) handling

[Pavan Balaji]

On 12/01/2010 07:59 PM, Yauheni Zelenko wrote:
> I checked nightly with modified cpi example and signals (SIGTSTP)
> are propagated correctly. However cpi was not suspended when running
> in MPI mode.
>
> Looks like SIGTSTP handling is more complicated then just relaying it
> to MPI processes...

Yes, it's harder for MPI processes as the MPI library itself will need 
to do some bookkeeping (maybe checkpointing, or at least resource 
cleanup). But for non-MPI applications this should work correctly.

  -- Pavan

-- 
Pavan Balaji
http://www.mcs.anl.gov/~balaji