Change replication in the cluster

[James McMahon <js...@gmail.com>]

We have been experiencing occasional failures to restart nifi services on a
cluster node because the flow.xml.gz falls out of synch with the other
nodes of the cluster. While researching this, found an article that
discusses possible causes. That article mentions this:

"A change replication request was made to all nodes in the cluster. One or
more nodes failed to process that request in the configured allowable
nifi.cluster.node.connection.timeout and/or nifi.cluster.node.read.timeout."

My question: is that change replication request exclusively node-to-node
communication, or does our external zookeeper play any role in that?

Thank you.