[
https://issues.apache.org/jira/browse/BAHIR-99?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16618437#comment-16618437
]
ASF GitHub Bot commented on BAHIR-99:
-------------------------------------
Github user meijies commented on the issue:
https://github.com/apache/bahir-flink/pull/17
currently, async kudu session flush for every piece of data. it will cause
poor performance(2w+/s). Should we flush micro batch data for streaming case? I
use 1s interval to flush the data, the performance is 40w+/s at my environment.
> Kudu connector to read/write from/to Kudu
> -----------------------------------------
>
> Key: BAHIR-99
> URL: https://issues.apache.org/jira/browse/BAHIR-99
> Project: Bahir
> Issue Type: New Feature
> Components: Flink Streaming Connectors
> Affects Versions: Flink-1.0
> Reporter: Rubén Casado
> Assignee: Joao Boto
> Priority: Major
> Fix For: Flink-Next
>
>
> Java library to integrate Apache Kudu and Apache Flink. Main goal is to be
> able to read/write data from/to Kudu using the DataSet and DataStream Flink's
> APIs.
> Data flows patterns:
> Batch
> - Kudu -> DataSet<RowSerializable> -> Kudu
> - Kudu -> DataSet<RowSerializable> -> other source
> - Other source -> DataSet<RowSerializable> -> other source
> Stream
> - Other source -> DataStream <RowSerializable> -> Kudu
> Code is available in https://github.com/rubencasado/Flink-Kudu
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
