We have been using Ironport for over three years now and have finally decided to tie it in with our Domino LDAP. Unfortunately, over the years we have been lenient with our mail policies and users have gotten used to using a number of email address variations. To capture these, we are searching on the following:

(|(mail={a})(email={a})(cn={u})(uid={u}))

"mail" is their normal address; "email" is an alternate for people who have had a name change due to divorce, marriage, etc.; "cn" is first initial-last name, and "uid" is last name-first initial.

This works fine using LDAP Browser, the Ironport LDAP tester, and a few limited-time live implementations. However, it is quite slow: 4-5 seconds per lookup. Taking the advice of someone in another post on this message board, we full text indexed our names.nsf file. The results were much faster, generally under a second.

However, the results also became more erratic. Though we didn't change our query string, we were now getting query failures when entering first_last@xyz.com (our standard address - the "mail" field). This only happens to users who don't have an "email" field in LDAP. If they have "mail" and "email" then the results come back positive.

Has anyone experienced this before while using Domino LDAP? I would hate to abandon full text indexing as it speeds the lookups considerably.