[ 
https://issues.apache.org/jira/browse/NIFI-11783?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17744770#comment-17744770
 ] 

ASF subversion and git services commented on NIFI-11783:
--------------------------------------------------------

Commit eb09df672b75cc65dbedbee363edc8fd715ba891 in nifi's branch 
refs/heads/support/nifi-1.x from Mark Payne
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=eb09df672b ]

NIFI-11783: Fixed a bug in which a FlowFile split into multiple and then 
removed did not properly handle retries when one of the outputs was retried.

Signed-off-by: Matt Burgess <[email protected]>


> Issue with rollback on retry connection
> ---------------------------------------
>
>                 Key: NIFI-11783
>                 URL: https://issues.apache.org/jira/browse/NIFI-11783
>             Project: Apache NiFi
>          Issue Type: Bug
>          Components: Core Framework
>    Affects Versions: 1.16.0, 1.17.0, 1.16.1, 1.16.2, 1.16.3, 1.18.0, 1.19.0, 
> 1.20.0, 1.19.1, 1.21.0, 1.22.0
>            Reporter: Pierre Villard
>            Assignee: Mark Payne
>            Priority: Critical
>             Fix For: 2.0.0, 1.23.0
>
>         Attachments: Bug_Lookup.json
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Consider a GenerateFlowFile -> LookupRecord flow.
> The GenerateFlowFile generates a flow file with two records.
> The LookupRecord processor is configured with matched/unmatched output 
> relationships.
> One of the record is matching an entry in the lookup service, the other 
> record does not match anything.
> With default configuration, the FF would be split into one flow file of one 
> record in the matched connection and one flow file of one record in the 
> unmatched relationship.
> This is expected behavior.
> Now, if I configure the retry mechanism (NIFI-9475) on the unmatched 
> relationship, no flow file is going into the unmatched relationship and I 
> have only one flow file with one record in the matched relationship. We're 
> effectively loosing the records that didn't match anything in the lookup 
> service.
> The expected behavior is to have all of the child flow files should be 
> removed and the original flow file should be penalized and rolled back.
> Find attached a flow definition to easily reproduce the problem.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to