There is no restriction to use the C/R functionality in Open MPI in a TM environment (that I am aware of), if you use the ompi-checkpoint/ompi-restart commands directly.
If you want TM to checkpoint/restart Open MPI processes for you as part of the resource management role, then there is a bit of a work around that you have to go through. The 'cr_mpirun' wrapper (mentioned in the email) that BLCR is/will be providing does the necessary things to make the two work together. The BLCR folks would be the best to contact if there are issues of compatibility when using that script since they maintain it. -- Josh On Feb 21, 2011, at 9:57 AM, Jeff Squyres wrote: > On Feb 21, 2011, at 12:50 AM, DOHERTY, Greg wrote: > >> blcr needs cr_mpirun to start the job without torque support to be able >> to checkpoint the mpi job correctly. > > Josh -- > > Do we have a restriction on BLCR support when used with TM? > > -- > Jeff Squyres > jsquy...@cisco.com > For corporate legal information go to: > http://www.cisco.com/web/about/doing_business/legal/cri/ > > _______________________________________________ > users mailing list > us...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/users > ------------------------------------ Joshua Hursey Postdoctoral Research Associate Oak Ridge National Laboratory http://users.nccs.gov/~jjhursey
