You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Haibo Chen (JIRA)" <ji...@apache.org> on 2018/10/22 17:08:00 UTC
[jira] [Updated] (YARN-8921) SnapshotBasedOverAllocationPolicy
always caps the amount of memory availabe to 4 GBs
[ https://issues.apache.org/jira/browse/YARN-8921?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Haibo Chen updated YARN-8921:
-----------------------------
Summary: SnapshotBasedOverAllocationPolicy always caps the amount of memory availabe to 4 GBs (was: SnapshotBasedOverAllocationPolicy incorrectly caps the amount of memory availabe in bytes to Integer.MAX_VALUE)
> SnapshotBasedOverAllocationPolicy always caps the amount of memory availabe to 4 GBs
> ------------------------------------------------------------------------------------
>
> Key: YARN-8921
> URL: https://issues.apache.org/jira/browse/YARN-8921
> Project: Hadoop YARN
> Issue Type: Sub-task
> Affects Versions: YARN-1011
> Reporter: Haibo Chen
> Assignee: Haibo Chen
> Priority: Major
>
> The memory overallocate threshold is a float, so is (overAllocationThresholds.getMemoryThreshold() *
> containersMonitor.getPmemAllocatedForContainers()). Because Math.round(float) return an int, this would cap effectively the amount of memory available for overallocation to Integer.MAX_VALUE, [see the code at here|https://github.com/apache/hadoop/blob/fa864b8744cfdfe613a917ba1bbd859a5b6f70b8/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/containermanager/scheduler/SnapshotBasedOverAllocationPolicy.java#L45]
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org