Skip to content

Commit 45bb853

Browse files
committed
"realm not local"
1 parent a700eae commit 45bb853

File tree

1 file changed

+20
-1
lines changed

1 file changed

+20
-1
lines changed

sections/errors.md

Lines changed: 20 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -554,4 +554,23 @@ attempts which may or may not be correct.
554554

555555
## Against Active Directory: `Realm not local to KDC while getting initial credentials`
556556

557-
Nobody knows.
557+
Nobody quite knows.
558+
559+
It's believed to be related to Active Directory cross-realm/forest stuff, but there
560+
are hints that it can also be raised when the kerberos client is trying to auth
561+
with a KDC, but supplying a hostname rather than the realm.
562+
563+
This may be because you have intentionally or unintentionally created [A Disjoint Namespace](https://technet.microsoft.com/en-us/library/cc731125(v=ws.10).aspx))
564+
565+
If you read that article, you will get the distinct impression that even the Microsoft
566+
Active Directory team are scared of Disjoint Namespaces, and so are going to a lot of
567+
effort to convince you not to go there. It may seem poignant that even the developers of
568+
AD are scared of this, but consider that these are probably inheritors of the codebase,
569+
not the original authors, and the final support line for when things don't work. Their
570+
very position in the company means that they get the worst-of-the-worst Kerberos-related
571+
problems. If they say "Don't go there", it'll be based on experience of fielding those
572+
support calls *and from having seen the Active Directory source code.*
573+
574+
575+
* [Kerberos and the Disjoint Namespace](http://www.networkworld.com/article/2347477/microsoft-subnet/kerberos-and-the-disjoint-namespace.htmla)
576+
* [Kerberos Principal Name Canonicalization and Cross-Realm Referrals](https://tools.ietf.org/html/rfc6806.html)

0 commit comments

Comments
 (0)