Comment 10 for bug 1443913

You should be able to take a look at the registry logs from the snapshot attached to the bug. Unfortunately from what I saw there wasn't anything out of the ordinary in the registry logs and the only error was in the API log. Since this is a test around controller failures and we're using memcache as our token store, I was wondering if this was caused by either a memcache connection error from a stale connection or perhaps an inconsistency around the token validation that occurs are the start of the v1 call and the internal calls that are made to process it. As far as I know, the token isn't technically expired it's just missing because of the controller failure scenario and the credentials are fine because you can wait a bit and the call will work just fine.