Known Issues Fixed in CDH 5.1.2
The following topics describe known issues fixed in CDH 5.1.2. See What's New in CDH 5.1.2 for a list of the most important upstream problems fixed in this release.
—Jobs can hang on NodeManager decommission owing to a race condition when continuous scheduling is enabled.
Workaround: Disable continuous scheduling by setting yarn.scheduler.fair.continuous-scheduling-enabled to false
— Sending a large amount of invalid data to the Thrift service can cause it to crash
Workaround: None. This is a longstanding problem, not a new issue in CDH 5.1.
— The metric ageOfLastShippedOp never decreases
This can cause it to appear as though the cluster is in an inconsistent state even when there is no problem.