[
https://issues.apache.org/jira/browse/HUDI-659?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476546#comment-17476546
]
Ethan Guo commented on HUDI-659:
--------------------------------
HoodieTimelineArchiveLog (new class name for HoodieCommitArchiveLog) is
engine-agnostic now. Closing this.
> Make HoodieCommitArchiveLog spark free
> --------------------------------------
>
> Key: HUDI-659
> URL: https://issues.apache.org/jira/browse/HUDI-659
> Project: Apache Hudi
> Issue Type: Task
> Components: Writer Core
> Reporter: vinoyang
> Assignee: vinoyang
> Priority: Major
>
> Currently, {{HoodieCommitArchiveLog}} depends on {{JavaSparkContext}} in its
> two methods: {{archiveIfRequired}} and {{getInstantsToArchive}}. These two
> methods pass {{JavaSparkContext}} to get {{HoodieTable}} object. After diving
> into the call chain, I found we can replace {{JavaSparkContext}} with
> {{Configuration}} and other some cleanup(e.g. HUDI-658). After that, we can
> make {{HoodieCommitArchiveLog}} spark free.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
