I have a new piece of infrastructure (a simple LDAP web service) that will be a drop-in replacement for something that currently exists in production.
The backend OpenLDAP server will not be upgraded at the same time.
In order to test the new infrastructure against the existing backend without risking corruption of current data, I would like to set up a secondary OpenLDAP server such that the pair operates in this way:
1) Any "write" request goes to the new "testing" LDAP server. 2) Any "read" request first goes to the existing "production" LDAP server, and then to the "testing" server if it fails.
Conceptually, this seems similar to a Master/Slave pair in which the testing node is the slave and the production node is the master, but I think that would require modification of the production node, which is a no-go.