[
https://issues.apache.org/jira/browse/HIVE-29307?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Denys Kuzmenko updated HIVE-29307:
----------------------------------
Fix Version/s: 4.2.0
(was: 4.3.0)
> Incorrect split calculation causing less container to launch
> ------------------------------------------------------------
>
> Key: HIVE-29307
> URL: https://issues.apache.org/jira/browse/HIVE-29307
> Project: Hive
> Issue Type: Bug
> Affects Versions: 4.0.0, 4.1.0, 4.0.1
> Reporter: Raghav Aggarwal
> Assignee: Raghav Aggarwal
> Priority: Major
> Labels: hive-4.2.0-must, pull-request-available
> Fix For: 4.2.0
>
>
> [https://github.com/apache/hive/blob/6b5febcc559e9f20ebea6fb04679477cf125144f/ql/src/java/org/apache/hadoop/hive/ql/exec/tez/HiveSplitGenerator.java#L363]
> {code:java}
> splits = inputFormat.getSplits(jobConf,
> numSplits.orElse(Math.multiplyExact(availableSlots, (int)waves))); {code}
> *Type casting of waves i.e. from float to int is causing less number of
> container launch*
>
> As good value of waves should be 1.7, 2.7, 3.7 etc as per
> [https://github.com/apache/tez/blob/5775337a5be3993d1aeb9b4c39134fe96c9b39f2/tez-mapreduce/src/main/java/org/apache/tez/mapreduce/grouper/TezSplitGrouper.java#L62]
> This incorrect calcutation will interfere with the
> org.apache.tez.mapreduce.grouper.TezSplitGrouper class.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
