[CRIU] CRIU <---> OpenMPI

[Pavel Emelyanov]

On 11/02/2013 04:31 AM, Christian Benvenuti (benve) wrote:
> Hello,
>   I have a question with regards to CRIU<--->OpenMPI.
> 
> I have noticed that in the CRIU TODO list [1] there is this item:
> 
> 	"Support one of MPI implementations  - worth starting with OpenMPI"
> 
> Is anyone actively working on this?

Right now -- no :( At least in Parallels.

> Is the current design (or idea/draft) based on an integration with the already existent
> checkpoint infrastructure present in OpenMPI [2] ?

I'm not an expert in OpenMPI architecture. So whether or not CRIU vs OpenMPI
would involve this facility or not is yet to be investigated.

I've only heard from several people, that work with OpenMPI that they have
plans to participate in CRIU development with this, but nothing more.

> In the recent LinuxCon/Plumbers conferences Pavel mentioned a few times the
> HPC periodic snapshot use case. Is this use case related to the above
> TODO item?

As I see it -- yes. When we do periodic snapshot we would have to take the
MPI context with us. But how should it look -- is to be found out.

> Thanks
> /Chris
> 
> [1] http://criu.org/Todo
> [2] http://www.open-mpi.org/~jsquyres/www.open-mpi.org/doc/v1.3/man1/ompi-checkpoint.1.php
> 

Thanks,
Pavel