Github user jtstorck commented on the issue:
https://github.com/apache/nifi/pull/3010
While testing the reconnection of a decommissioned node, with ~24,000 files
split between the two nodes, an error occurred after reconnecting the
decomissioned node and attempting to drop all the flowfiles in the queue, from
both nodes:
```
2018-09-20 15:25:22,594 ERROR [Drop FlowFiles for Connection
cbbf2971-0165-1000-ffff-ffff94b81269]
o.a.n.c.q.c.SocketLoadBalancedFlowFileQueue Failed to drop FlowFiles for
org.apache.nifi.controller.queue.clustered.SocketLoadBalancedFlowFileQueue@469caf69
java.lang.IllegalArgumentException: null
at org.apache.nifi.controller.queue.QueueSize.<init>(QueueSize.java:31)
at org.apache.nifi.controller.queue.QueueSize.add(QueueSize.java:67)
at
org.apache.nifi.controller.queue.clustered.SocketLoadBalancedFlowFileQueue.adjustSize(SocketLoadBalancedFlowFileQueue.java:514)
at
org.apache.nifi.controller.queue.clustered.SocketLoadBalancedFlowFileQueue.dropFlowFiles(SocketLoadBalancedFlowFileQueue.java:903)
at
org.apache.nifi.controller.queue.AbstractFlowFileQueue$2.run(AbstractFlowFileQueue.java:285)
at java.lang.Thread.run(Thread.java:748)
```
The decommission operation was successful, all flowfiles were moved from
the decommissioned node to the other node. After reconnecting the
decommissioned node, I couldn't clear the flowfile queue.
After restarting the cluster (both nodes), the queue showed as empty.
---
