When you ask the system to get you a ticket to host/foo.bar.com and it fails to get a ticket for various reasons, it puts the SPN in what's called a negative cache. Every time you ask for a ticket to an SPN the first thing the system does is check this negative cache and asks, "have I tried this recently and if so, did it fail?" If it failed, then it knows don't bother trying and short circuit the same error. This cache is valid for a certain period of time and is controlled by the "SpnCacheTimeout" value. There is a separate cache controlled by FarKdcTimeout that does exactly the same thing, except for when it can't contact a given KDC.
0 commit comments