[NIFI-1900] NiFi allows connections to be moved between processors in cluster when it should not

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.6.1
• Fix Version/s: 1.0.0, 1.0.0-Beta
• Component/s: Core Framework
• Assignee: Unassigned
• Reporter: Matthew Clarke
• Created: 19/May/16 13:08
• Updated: 04/Aug/16 21:23
• Resolved: 22/Jun/16 16:42

Description

• • • The result of this bug is one or more dropped nodes from the NiFi cluster requiring flow.xml.gz removal and restart of any dropped nodes.

Scenario that causes bug:
1. NiFi multi-node cluster setup.
2. Active flow between two processors.
3. Destination processor has actively running threads.
4. Stop destination processor.
5. Destination processor still has active threads after stop but not on every node.
6. Drag processor connection to new destination processor.
7. Those nodes where the original destination processor did not have active threads successfully move connection.
8. Those nodes with hung or active threads fail to move connection and are then dropped from cluster because of flow no longer matches what is on NCM.

Recovery requires remove of flow.xml.gz on all dropped nodes and restart of those nodes.

Issue Links

• links to: GitHub Pull Request #550