Re: Upgrading OpenLDAP, problem with data (no structural object class).

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 01:18 AM 2/14/2005, Andrew McCall wrote:
>Hi Folks,
>
>In upgrading my OpenLDAP to an earlier version, I have realised how
>"incorrect" the data is in it, and am now sifting through about 3000
>entries to try and stuff the contents back in it!  One of the problems
>I am having is where I have the following type of entry:
>
>dn: cn=Web Content Administrators,o=oldhammbc, c=uk
>cn: Web Content Administrators
>objectClass: top
>objectClass: qmailUser
>qmailUID: 503
>qmailGID: 507
>accountStatus: active
>deliveryMode: forwardonly
>mail: XX.XXX.XXXXXXX@oldham.gov.uk
>mailForwardingAddress: XX.XXX.XXXXXXX@oldham.gov.uk
>uid: XX.XXX.XXXXXXX
>subschemaSubentry: cn=Subschema
>
>This is incorrect as it has no structural object class and I am
>getting the following error when I try to add the entry "slapadd:
>dn="cn=Web Content Administrators,o=oldhammbc,c=uk" (line=38): (65) no
>structural object class provided".  I don't need any of the data in inetOrgPerson, or organizationalPerson, so how can I make accounts
>like this go into the LDAP?

Please note that slapd(8) could care less which structural
object class you provide (as long as it has been provided with an
appropriate specification of the class).  Hence, your questions
here are not specific to OpenLDAP Software.  Your questions
are specific to the directory applications you are deploying.

Seems you consider this object an account, maybe should add an
structural object class intended to represent an account.  But
I suggest you consult the documentation of the application
software you are using to determine which structural object
classes are appropriate and then select one of these.  Specifically,
I suggest you ask qmail folks (on a qmail list) which structural
object classes their auxiliary object class is intended to be
attached to.

>Can I change the qmailUser schema so that it is a structural
>object class?

It is generally inappropriate to modify published schema elements.

Kurt