The secondary-RM-to-primary-RM (and NM) redirect issued by YARN in a kerberoized cluster is not a proper (http-location-301-style) redirect, thus Ambari - which uses curl - does not find the "right" node. This, in turn, is triggering alerts in Ambari.
A network dump of the ambari poll against the secondary RM looks like:
GET /jmx?qry=Hadoop:service=ResourceManager,name=RMNMInfo HTTP/1.1
HTTP/1.1 200 OK
Refresh: 3; url=http://
This is standby RM. Redirecting to the current active RM:
Comment from Jonathan Hurley email@example.com -
1) There should be a ticket filed for YARN to have their HA mode use a proper redirect
2) Ambari might not want to produce an UNKNOWN response here since it gives a false feeling that something went wrong.
I've also filed AMBARI-12995 with the ambari team.