Klavdiya Popova created NIFI-15813:
--------------------------------------
Summary: ExecuteSQL processors don't transfer/create flowfile for
failure when output batch size > 0
Key: NIFI-15813
URL: https://issues.apache.org/jira/browse/NIFI-15813
Project: Apache NiFi
Issue Type: Bug
Components: Extensions
Reporter: Klavdiya Popova
Attachments: image-2026-04-10-15-03-45-563.png
I have noticed illogical behavior in ExecuteSQL processors when handling errors
with output batch size > 0. It can throw an exception without transferring any
files to the failure relationship.
In my case Max rows per flow file > 0.
!image-2026-04-10-15-03-45-563.png!
The processor consumes the input flow file and drops it as soon as the first
flow file is sent to the success relationship. At the same time the processor
continues to fetch data records and send them in new flow files to the output
queue. If an error occurs after dropping the original flow file, the processor
produces the error log and doesn't send any flow files to the failure
relationship.
I would expect the processor either to create a flow file as a child to the
original file and send it to the failure relationship, or to keep the original
flow file until all rows are fetch and in the case of an exception to transfer
it to the failure relationship.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
