[
https://issues.apache.org/jira/browse/HUDI-5005?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
JinxinTang updated HUDI-5005:
-----------------------------
Description:
# When stream write reuse aborted instant, there is chance this one is older
than instant of coordinator
# For org.apache.hudi.sink.common.AbstractStreamWriteFunction#instantToWrite,
there is chance to get two different instants(the older and the new one), but
only send the older to coordinator, so the coordinator think the new one is
dirty and delete it, which will cause FileNotFoundException.
> flink stream write reuse abort instant will lead to coordinator delete file
> not right.
> --------------------------------------------------------------------------------------
>
> Key: HUDI-5005
> URL: https://issues.apache.org/jira/browse/HUDI-5005
> Project: Apache Hudi
> Issue Type: Bug
> Reporter: JinxinTang
> Assignee: JinxinTang
> Priority: Major
> Labels: pull-request-available
>
> # When stream write reuse aborted instant, there is chance this one is older
> than instant of coordinator
> # For
> org.apache.hudi.sink.common.AbstractStreamWriteFunction#instantToWrite, there
> is chance to get two different instants(the older and the new one), but only
> send the older to coordinator, so the coordinator think the new one is dirty
> and delete it, which will cause FileNotFoundException.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
