Implementing Attribute Value Uniqueness
Some attribute values ought to remain unique. If you are using
uid values as RDNs to distinguish between millions of
user entries stored under ou=People, then you do not
want your directory too contain two or more identical
uid values. If your credit card or mobile number is
stored as an attribute value on your directory entry, you certainly do not
want to share that credit card or mobile number with another customer.
The same is true for your email address.
The difficulty for you as directory administrator lies in
implementing attribute value uniqueness without sacrificing the high
availability that comes from using OpenDJ's loosely consistent,
multi-master data replication. Indeed OpenDJ's replication model lets
you maintain write access during network partitions for directory
applications. Yet, write access during a network partition can result in the
same, theoretically unique attribute value getting assigned to two different
entries at once. You do not notice the problem until the network partition
goes away and replication resumes.
This chapter shows you how to set up attribute value uniqueness
in your directory environment.