[
https://issues.apache.org/jira/browse/NIFI-5989?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16763882#comment-16763882
]
Mark Payne commented on NIFI-5989:
----------------------------------
[~goosalex] thanks for the contribution! I was able to test this and verify
that it works as expected. This is definitely a good distinction to make,
Record Count vs. FlowFile count per batch. I did update the names of the
properties slightly to clarify the difference between the two properties but
otherwise all checks out and is good. Thanks again! +1 merged to master.
> Improve PutKudu BatchSize handling
> ----------------------------------
>
> Key: NIFI-5989
> URL: https://issues.apache.org/jira/browse/NIFI-5989
> Project: Apache NiFi
> Issue Type: Improvement
> Components: Extensions
> Reporter: Alex Goos
> Priority: Major
> Labels: kudu, nifi
> Fix For: 1.9.0
>
> Attachments:
> 0001-NIFI-5989-PutKudu-Additional-FF-Queue-length-setting.patch
>
>
> Current "Batch size" property of PutKudu affects both: the number of
> Flowfiles pulled per OnTrigger and the size of the Kudu client modification
> buffer.
> If the Flowfiles contain a considerable amount of records, then a
> disproportionate amount of data is pulled in and deserialized into memory,
> when in AUTO_FLUSH_BACKGROUND mode.Â
> We propose introducing a separate setting for the batch size of FlowFiles.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
