There are however issues when using a replicated OID deployment. OIM uses a process called LDAPSync which is used to keep the entries inside its internal database in sync with the entries in LDAP.
To prevent the process having to process every entry in LDAP each time, it uses a changelog which allocates a change number to every transaction in the directory. If you have an OID replicated environment, then the change numbers
in each OID cluster are different.
in each OID cluster are different.
This causes issues for failover. You can alleviate this by:
1. Ensuring LDAPSync (OIM reconciliation jobs) only run against a single OID cluster.
2. If you do need to failover to a second OID cluster, then you will need to:
a. Disable the incremental OIM reconciliation tasks.
b. Run a full Reconciliation against the new OID cluster.
c. Update the OIM change number to reflect that of the new OID cluster.
d. Re-enable incremental OIM reconciliation tasks.
This is not necessary for OUD based solutions which use a cookie based changelog.
