"Fail back" behavior of the Active Directory KCC when using redundant higher cost site links to a DR site

Solution 1:

The KCC will fail back automatically. It runs every 15 minutes as you know, and it performs "both phase 1 and phase 2" to use your example above, as part of its logic on each run.

Essentially, the KCC will keep trying the "failed" DC indefinitely, just in case it comes back. When it does come back, you'll see this event in your log:

Event 1129

Source: ActiveDirectory_DomainService

To improve the replication load of Active Directory Domain Services, a replication connection from the following source directory service to the local directory service was deleted.

[The failover connections that were automatically created when your main DC failed will be here.]

Here's a good MS blog post with a lot more detail. It's a 3 parter, very detailed, I linked part 3 that mentions fail back: http://blogs.technet.com/b/isrpfeplat/archive/2011/12/11/disaster-recovery-site-and-active-directory-part-3-of-3.aspx