Various proposals have sought to increase data reliability and availability on cluster networks. Database systems have their own protection schemes for failover conditions so that they can continue to serve their applications and user processes. However, HA (High-Availability) requires a high cost in that it uses a complex code in the DBMS. This also necessitates an additional setup procedure for database management and administration, as well as for the setup of some specialized hardware. In this paper, the authors present what is termed the \"Selective Replication Manager\" (SRM), which interacts with the system status of the Linux operating system.