RMD should reapply the policies upon restart of RMD

[Shihta]

Here is a scenario cause RMD loses established policies

1. Run RMD in a fresh environment, there is no workload/COS in RMD/resctrl
2. Create a workload via POST /workloads
3. Kill RMD process, umount resctrl
4. mount resctrl, Run RMD again
5. Query workloads via GET /workloads,
   and then you'll see there is no any workload in response.