[NIFI-5866] Cluster gets out of sync: "X is not the most up-to-date revision."

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Duplicate
• Affects Version/s: 1.6.0
• Fix Version/s: 1.13.0
• Environment: CentOS 6.9, NiFi 1.6.0 cluster of 4, running on AWS EC2 VMs.
• Assignee: Unassigned
• Reporter: Joe Gresock
• Created: 04/Dec/18 16:58
• Updated: 23/Nov/20 14:36
• Resolved: 23/Nov/20 14:36

Description

Since upgrading to 1.6.0, I have started to see this error sometimes when I try to modify a component (start, stop, configure, move, etc.):

 

Node X is unable to fulfill this request due to: [12, <some uuid>, <some uuid>] is not the most up-to-date revision.  This component appears to have been modified.

 

The only way I have found to resolve this is to disconnect and then delete the node from the cluster, then try again, repeating the process until the flow can be modified.  Afterwards, I have to stop nifi on the disconnected nodes, delete their flow.xml.gz, and let them rejoin the cluster with a fresh copy of the flow.

 

This sometimes happens multiple times per day.

Issue Links

• duplicates: NIFI-7920 When enabling Controller Services, if Cluster Coordinator fails to enable, can result in service becoming unmodifiable with error "Revision ... is not the most up-to-date revision. This component appears to have been modified"