Stephen O'Donnell

We are refactoring how the container checks are handled in the new RM in #3743, so once that goes in this one will need changed to work in the way...

We have left this new handler out of the RM chain deliberately for now, as its not really complete without the deleting handler. The plan is to implement the deleting...

I think it is reasonable to remove containers that are empty. There isn't a good reason to keep them beyond debugging and if a DN comes back up with a...

IMO I think this change looks good. Perhaps we should add one more test that confirms the reference is not deleted if the config flag is false, just to catch...

I can kind of see this both ways. If we take it to one extreme - if we want to ensure stuff is kept around for debugging, why are we...

I'm not sure about this change. I feel it adds quite a bit of complexity to the model to handle what is really an edge case for a small cluster....

I feel we should handle this in the under-replication handler. In normal circumstances, the container can be under-replicated and it might be over replicated too, but if we can create...

There is some work going on to fix this sort of issue in different ways. #3482 is going to limit the number of replications that can be inflight at any...

The changes look good, but I think it would be much more useful if we could track metric at the decommissioning node level too. Ie: ``` TotalTrackedContainersUnderReplicatedForHostname = xyz ```...

HDFS has a metric like this: ``` "DecomNodes" : "{\"cdh-6x-of-1.cdh-6x-of.root.hwx.site:20002\":{\"xferaddr\":\"172.27.52.133:20002\",\"underReplicatedBlocks\":0,\"decommissionOnlyReplicas\":0,\"underReplicateInOpenFiles\":0}}", ``` It seems to register a MBean instance in the FSNameSystem class. Then it has a few places it provides...