Solved bug in select/cons_res plugin which makes jobs requesting

tasks_per_node and cpus_per_task not getting proper allocation and
failing in srun with "Requested node configuration is not	available"