You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Rakshith C (Jira)" <ji...@apache.org> on 2023/02/07 05:24:00 UTC
[jira] [Resolved] (HIVE-26937) Batch events during incremental replication to avoid O.O.M
[ https://issues.apache.org/jira/browse/HIVE-26937?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Rakshith C resolved HIVE-26937.
-------------------------------
Resolution: Fixed
> Batch events during incremental replication to avoid O.O.M
> ----------------------------------------------------------
>
> Key: HIVE-26937
> URL: https://issues.apache.org/jira/browse/HIVE-26937
> Project: Hive
> Issue Type: Improvement
> Components: Hive
> Reporter: Rakshith C
> Assignee: Rakshith C
> Priority: Major
> Labels: pull-request-available
> Time Spent: 1h
> Remaining Estimate: 0h
>
> * Currently incremental replication flow of hive dumps all events read from notification logs sequentially in staging directory.
> * Repl Load loads all the event directories present in staging directory to a list and processes them.
> * This has caused O.O.M issues when number of events are large.
> Hence introducing batching of events where Repl Dump dumps events in batches and Repl Load processes events batch by batch to avoid O.O.M.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)