[
https://issues.apache.org/jira/browse/YARN-8391?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16517335#comment-16517335
]
Miklos Szegedi commented on YARN-8391:
--------------------------------------
Thank you for the patch [~snemeth]. Is the unit test does not seem to be
related, is it?
> Investigate AllocationFileLoaderService.reloadListener locking issue
> --------------------------------------------------------------------
>
> Key: YARN-8391
> URL: https://issues.apache.org/jira/browse/YARN-8391
> Project: Hadoop YARN
> Issue Type: Bug
> Components: fairscheduler
> Affects Versions: 3.2.0
> Reporter: Haibo Chen
> Assignee: Szilard Nemeth
> Priority: Critical
> Attachments: YARN-8191.001.patch, YARN-8391.002.patch
>
>
> Per findbugs report in YARN-8390, there is some inconsistent locking of
> reloadListener
>
> h1. Warnings
> Click on a warning row to see full context information.
> h2. Multithreaded correctness Warnings
> ||Code||Warning||
> |IS|Inconsistent synchronization of
> org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.AllocationFileLoaderService.reloadListener;
> locked 75% of time|
> | |[Bug type IS2_INCONSISTENT_SYNC (click for
> details)|https://builds.apache.org/job/PreCommit-YARN-Build/20939/artifact/out/new-findbugs-hadoop-yarn-project_hadoop-yarn_hadoop-yarn-server_hadoop-yarn-server-resourcemanager.html#IS2_INCONSISTENT_SYNC]
>
> In class
> org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.AllocationFileLoaderService
> Field
> org.apache.hadoop.yarn.server.resourcemanager.scheduler.fair.AllocationFileLoaderService.reloadListener
> Synchronized 75% of the time
> Unsynchronized access at AllocationFileLoaderService.java:[line 117]
> Synchronized access at AllocationFileLoaderService.java:[line 212]
> Synchronized access at AllocationFileLoaderService.java:[line 228]
> Synchronized access at AllocationFileLoaderService.java:[line 269]|
> h1. Details
> h2. IS2_INCONSISTENT_SYNC: Inconsistent synchronization
> The fields of this class appear to be accessed inconsistently with respect to
> synchronization. This bug report indicates that the bug pattern detector
> judged that
> * The class contains a mix of locked and unlocked accesses,
> * The class is *not* annotated as javax.annotation.concurrent.NotThreadSafe,
> * At least one locked access was performed by one of the class's own
> methods, and
> * The number of unsynchronized field accesses (reads and writes) was no more
> than one third of all accesses, with writes being weighed twice as high as
> reads
> A typical bug matching this bug pattern is forgetting to synchronize one of
> the methods in a class that is intended to be thread-safe.
> You can select the nodes labeled "Unsynchronized access" to show the code
> locations where the detector believed that a field was accessed without
> synchronization.
> Note that there are various sources of inaccuracy in this detector; for
> example, the detector cannot statically detect all situations in which a lock
> is held. Also, even when the detector is accurate in distinguishing locked
> vs. unlocked accesses, the code in question may still be correct.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: yarn-issues-h...@hadoop.apache.org
