You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-issues@hadoop.apache.org by "guangyi lu (JIRA)" <ji...@apache.org> on 2018/12/11 12:46:00 UTC
[jira] [Updated] (HDFS-14074) DataNode runs async disk checks
maybe throws NullPointerException, and DataNode failed to register to
NameSpace.
[ https://issues.apache.org/jira/browse/HDFS-14074?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
guangyi lu updated HDFS-14074:
------------------------------
Attachment: HDFS-14074-latest.patch
> DataNode runs async disk checks maybe throws NullPointerException, and DataNode failed to register to NameSpace.
> ------------------------------------------------------------------------------------------------------------------
>
> Key: HDFS-14074
> URL: https://issues.apache.org/jira/browse/HDFS-14074
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs
> Affects Versions: 2.8.0, 3.0.0
> Environment: hadoop-2.7.3, hadoop-2.8.0
> Reporter: guangyi lu
> Priority: Major
> Labels: HDFS, HDFS-11114
> Fix For: 2.8.0, 2.7.3, 3.0.0
>
> Attachments: HDFS-14074-latest.patch, HDFS-14074.patch, WechatIMG83.jpeg
>
> Original Estimate: 48h
> Remaining Estimate: 48h
>
> In ThrottledAsyncChecker class，it members of the completedChecks is WeakHashMap, its definition is as follows：
> this.completedChecks = new WeakHashMap<>();
> and one of its uses is as follows in schedule method:
> if (completedChecks.containsKey(target)) {
> // here may be happen garbage collection，and result may be null.
> final LastCheckResult<V> result = completedChecks.get(target);
> final long msSinceLastCheck = timer.monotonicNow() - result.completedAt;
> 。。。。
> }
> after "completedChecks.containsKey(target)"， may be happen garbage collection， and result may be null.
> the solution is：
> this.completedChecks = new ReferenceMap(1, 1);
> or
> this.completedChecks = new HashMap<>();
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-help@hadoop.apache.org