Matthew Clarke created NIFI-1297:
------------------------------------
Summary: NiFi Cluster should persist list of all connected nodes
and not just the primary node
Key: NIFI-1297
URL: https://issues.apache.org/jira/browse/NIFI-1297
Project: Apache NiFi
Issue Type: Improvement
Components: Core Framework
Affects Versions: 0.4.0
Reporter: Matthew Clarke
Priority: Minor
Currently with an active NiFI cluster the only node that is persisted through a
NCM reboot is the primary node. We should persist all previously connected
nodes through a NCM reboot.
The problem is that in most user cases the Admin who would restart the software
and the DFMs who manage the flows are different people. Not persisting all
nodes give a DFM the ability to alter the flow before all nodes have
reconnected.
I believe the original intend was to not have any dependency on all nodes or
even the same nodes reconnect; however, I feel the more typical case is the
same nodes rejoining. It is more beneficial for the DFMs to see what nodes have
failed to join and take appropriate actions. That action may be to drop the
node manually or go investigate why it did not join as expected.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
