Zombie Nifi instance detection

[#1]

Nifi version 1.1.0.2.1.0.0-165. Two single node instances, S1 and S2 running on same VM. S1 has output ports at the top level. S2 defines a RPG referring back to S1's output port. S2 runs system out of disk space. S2 is brought down, logs cleared, and setting changed to only keep one day of logs (this is development environment). S2 is brought back up, everything seemed fine. Later what's noticed is S2 RPG to S1 starts showing connection errors. Did a status check on S1 and it showed S1 was still running. S1's UI came up fine also. When I looked closer though I found logging on S1 had stopped where the last message was

[2017-02-21 10:59:06,413 ERROR [pool-8-thread-1] org.wali.MinimalLockingWriteAheadLog Failed to create new journal for Partition-74, java.io.IOException: No space left on device] due to {} java.io.IOException: No space left on device]

Was just wondering if there was some other way to detect S1 had become a zombie other than the UI error being shown on S2's RPG? I cannot be definite but I assume if I had tried to really do anything on S1's UI it would have failed (though it came up ok).

Thanks

[#2]

You need to set them up in a Zero master cluster

Zookeeper

http://docs.hortonworks.com/HDPDocuments/HDF2/HDF-2.1.1/bk_dataflow-administration/content/embedded_...

Node Failure

http://docs.hortonworks.com/HDPDocuments/HDF2/HDF-2.1.1/bk_dataflow-administration/content/cluster-n...

Get email notifications

http://docs.hortonworks.com/HDPDocuments/HDF2/HDF-2.1.1/bk_dataflow-administration/content/notificat...