[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: still unclear on error 69

Tony Earnshaw wrote:
Jon Roberts wrote:
If the server were down, the authentication failed, the user didn't have privileges to the data, etc.... there would've been a different error code telling me so. I think the 69 error is telling me something new and more specific, and I'd like to get to the bottom of it.

What it is telling you is, you have to have *all* the objectclasses necessary in the hierarchy before it can add what you want. See my last answer. 

I read your post. Did you read mine?

I'm only using top, person, organizationalperson, and inetorgperson.
What's missing? I understood your point about conflicts in strucutural objectclasses, but it doesn't apply.

The 69 error occurs when I attempt a modify operation on the
objectclass attribute to go from a [top, person] entry to a [top,
person, organizationalperson] or [top, person, organizationalperson, inetorgperson] entry.

Again the eternal premise: "If it works for 1,000 others, why doesn't 
it work for me?"

The only testimony I've heard for doing such an operation is from
another person who got the exact same error.

http://www.openldap.org/lists/openldap-software/200307/msg00644.html

Try it yourself. Assume and entry:

dn: cn=Mama, ou=People, o=family.org
objectclass: top
objectclass: person
cn: Mama
sn: Jones

Then try to implement the LDIF:

dn: cn=Mama, ou=People, o=family.org
changetype: modify
add: objectclass
objectclass: organizationalperson

And with ldapmodify you will get:

modifying entry "cn=Mama, ou=People, o=family.org"
ldapmodify: update failed: cn=Mama, ou=People, o=family.org
ldap_modify: Cannot modify object class (69)
additional info: structural object class modification from 'person' to 'organizationalperson' not allowed

Looks like Mama has to stay in the kitchen :(

I tried this on an OpenLDAP 2.1.22 server with a BDB backend I built and installed this morning. My question (rephrased) still stands: is there any way to add valid structural objectclasses to an existing entry that already has a strucutural objectclass through the protocol?

Jon Roberts
www.mentata.com