Top three reasons for implementing OID Replication:
1. Local Accessibility and Performance requirement. --- All major enterprise operate on geographically separated location with common directory and remote location are interconnected with low bandwidth its good idea to move to OID replication based on geographical locations for improved performance and local accessibility
2. High Availability & fault tolerance. --- Implementing Directory Server on two or more replicated (multimaster) servers provide High Availability and better fault tolerance.
3. Better return on Investment. ---Deploying replicated OID in two or more small servers is always better in terms of Return of Investment (Scalable to increasing user load)
Type of replication (Single Master, Multi Master, Fan-Out) Single Master: Only one supplier (Users can edit entry) and many consumers (Users can only read but not edit). Use LDAP or DB advanced replication protocol Multi Master: Replica OID are Supplier as well consumer at same time from other Replica OID. Using DB Advanced replication only (can't be achieved by LDAP based replica) Fan Out: A supplier replicating one consumer which in turn is updating other consumers.
Three command line tools which you can use to monitor and troubleshoot OID replication are
OID Replication Management Tools remtool: Replication Environment Management Tool oidreconcile:OID Reconciliation Tool hiqretry: Human Intervention Queue Manipulation Tool
Apart from this you can use GUI tool i.e. oidadmin to administer Member of Directory Replication Group
Key Points you should know for OID Replication: OID Replication Architecture is based on loose connectivity model (Two Replica Directory are not guaranteed to be consistent in real time)
In my coming posts I'll explain architecture of Multi Master Replication with processes at both Supplier and Consumer side (All Nodes in Multi Master replication act as both consumer and supplier)