> On 26 Mar 2015, at 16:01 , Ralph Castain <r...@open-mpi.org> wrote: > >> >> On Mar 26, 2015, at 1:33 AM, Mark Santcroos <mark.santcr...@rutgers.edu> >> wrote: >> >> Hi guys, >> >> Thanks for the follow-up. >> >> It appears that you are ruling out that Munge is required because the system >> runs TORQUE, but as far as I can see Munge is/can be used by both SLURM and >> TORQUE. >> (http://docs.adaptivecomputing.com/torque/4-0-2/Content/topics/1-installConfig/serverConfig.htm#usingMUNGEAuth) > > Not really ruling it out, Mark, but I didn’t consider it likely because then > munge would indeed have to be on the compute nodes. Otherwise, the sister > moms wouldn’t be able to authenticate back to the mom on the IO node. > > To be clear, I’m not 100% sure what is using munge on the IO node. My real > point was only that there are other subsystems that can use such security > services, and that those subsystems might not extend into the compute node > itself. Thus, the need to work in multiple security domains is going to exist > into the future.
Right, I think I'm clear on the issue now :-)
