When you enable field engineering (debug) logging to trace an LDAP query, the following event log shows that the LDAP query is an inefficient query:Note The attributes that are used in this event are only examples.
Additionally, you experience high CPU utilization and a slow response time. If the database is considerably bigger than physical memory available to the directory server, you may also see increased disk IO while the processing such a query.
When you inspect the attributes in the search filter, you find that they all have indexes that are defined. And if attributes do not have indexes that are defined, and you add the indexes through a schema change, the problem persists or does not improves much.
The LDAP server can only use one index while processing a paged query. This is because the LDAP implementation for paged searches does not create an expensive context for the query and thus only uses one index for a paged query.
Noe that by default, paged queries are enabled for some LDAP client libraries. Therefore, you may have to write additional code in your application to enable and disable paged queries as appropriate for your specific situation.
Article ID: 951581 - Last Review: Nov 9, 2010 - Revision: 1