There are two types of servers defined in the ClusterReplica environment: the “master” and “replica”. While the “master”
is the data replication source, the “replica” is the data replication destination. Some terminologies involved using
ClusterReplica Enterprise are introduced here.
The Primary master – this is a data replication source server that is also the manager of the configuration of the
entire clustering structure. It defines other data replication source servers and the replication destination servers. It
is the only server that can be failed over to a Secondary station for server-availability needs.
The repli-master – this is a data replication source server that needs to replicate the data files on it to other
servers in local or remote locations. While functioning mainly as the data replication source, a repli-master can
also be used as a replica member to receive data from other master servers. It is different from the Primary
master in the way that it does not have the management power on the clustering.
The replica – this is a data replication destination server. It takes in data from the master servers and manages
the data in MS DPM style for easy data distribution and recovery.
The Secondary station – this is a replica station that besides function as the data replication destination, is also
defined as the Secondary station which is ready to take over the server services of the Primary masters should it
fails at any time. There can be only one Secondary station defined in a ClusterReplica environment. For a smooth
failover, this system is expected to have the similar system capacity of the Primary master and with identical
application and file structure.
In this example, there are three “master” stations: one Primary master, two repli-master(s) and three “replica” stations.
One of the “replica” station is functioning as the Secondary station. The “replica” stations receive and manage new data in
the manner of Microsoft’s DPM system so that when file-retrieving requests are processed, data files can be easily found
and restored. Also like the MS DPS, this file managing and retrieving process is entirely end-user self-service.
With this arrangement, all server malfunctions can be avoided:
