the front-end node that job was running on is no longer in slurm.conf

Fix issue when updating database for clusters that were previously deleted before upgrade to 2.2 database.

 invalid dim count in the RPC treated as an unpack error instead of aborting slurmctld
 the default dim size was being loaded from slurmdb_setup_cluster_flags(), not 
 slurmdb_setup_cluster_dims()

    IDLE (in case the previous node state was DOWN).

added new flag HAVE_CRAY_EMULATION for marking objects with emulation instead of a real cray machine