SIGTSTP propagation for 0.80 ?

[David Golden]

On 2005-09-20 08:48:49 +0200, Roy Dragseth wrote:
> >
> > I don't know if this will be possible, as Torque has to deal with cases
> > (such as MPICH and some commercial MPI apps, not to mention non-MPI

> I agree, I don't think it will be feasible to change the way torque handles 
> signals as this will be a quite deep change in the way torque behaves.  

You+Chris are right, of course. Though I've been looking through
the torque sources and IMHO making signal propagation behaviour a tunable
parameter per-job mightn't really be all that difficult after all, just a bit 
fiddly. Whether it's worthwhile is another matter, of course.

> mpich/p4 with mpiexec (0.78 with patch works, 0.80 don't?)

Well, I haven't yet had the chance to test the 0.80.tstp1 patch 
with simply a longer time between the TSTP and STOP  - It's quite
suspicious that qsig -s TSTP works and qsig -s suspend doesn't.