10. Distributing slapd DATA

For many sites, running one or more slapds that hold an entire subtree of data is sufficient. But sometimes it may be desirable to have one slapd refer to other slapds for a certain part of the tree. This can be accomplished by creating a referral entry in one slapd's database pointing to another slapd. For those familiar with X.500, a slapd referral entry is similar to an X.500 knowledge reference.

The referral entry acts as a mount point, glueing two slapd databases together. A referral entry has an objectclass of "referral" and is named by a ref attribute containing a URL pointing to the slapd holding the data below the mount point. This mechanism is very general and allows slapd databases that are not normally hierarchical to be grafted together.

An example should help illustrate things. Suppose your company is running a slapd and just purchased a new company, also running a slapd. You can easily connect the two databases by creating an entry like this in your slapd's database.

dn: ref="ldap://new.host/o=New Company,c=US", o=Your
company, c=US
objectclass: referral

Now any subtree search that has this entry in its scope will return a referral to the new company, in addition to any entries matched in your database. Referral-aware clients will continue the search at the new company's server.

A mechanism similar to this is used to support distributed indexing, described in Appendix C.

[View Next Section] [View Previous Section] [Return to Table of Contents]

Send comments about this page to: ldap-support@umich.edu