I have a Tiger Server, which is also functioning as a desktop workstation. Under Tiger operation the local users have records in the server’s Open Directory database. Using the Directory Access utility of Tiger, the server is bound to its own OD domain. Since obtaining a Leopard server running on a different machine, I have attempted to replace the binding of the Tiger Server machine to the new Leopard OD domain. This has failed, in that after successful setup of the bind log-ins using the user credentials from the Leopard OD domain don’t work.

The server has the LDAP search base dc=xserve,dc=a,dc=b,dc=net. The Leopard Server is an Open Directory Master. I have enabled authenticated binding (in Open Directory->Policy->Binding) and am requiring authenticated binding between directory and clients.

In the Tiger Server’s Directory Access utility I have temporarily unchecked the “Enable” checkbox binding the Tiger Server to itself. Instead I have created and enabled a new LDAP search policy to the server xserve.a.b.net. The LDAP mapping is set to “Open Directory Server” and the search base suffix is set to “cn=config,dc=xserve,dc=a,dc=b,dc=net”. The Authentication has “/LDAPv3/xserve.a.b.net” included in the list of directory domains as a “Custom path.” Said directory domain is listed ahead of “/LDAPv3/127.0.0.1”.

As a troubleshooting step I bound my MacBook Pro, which runs Leopard to the Leopard Server domain, successfully, with the ability to log into a session hosted by the Leopard Server.

So, what’s the difference between Tiger and Leopard as a client to a Leopard-hosted Open Directory domain?

Dominik Hoffmann