[
https://issues.apache.org/jira/browse/HDFS-14037?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16771760#comment-16771760
]
Takanobu Asanuma edited comment on HDFS-14037 at 2/19/19 9:39 AM:
------------------------------------------------------------------
[~brahmareddy] Sorry for late response.
* Considering that a new instance of RouterWebHdfsMethods is always created
every when a WebHDFS API is executed, storing the factory as an instance filed
may be no meaning.
* We can implement a method in DFSck to destroy the factory. But it may be
overkill because DFSck is a stand alone command, hadoop fsck, the lifetime is
short, and the factory is created only once when the command is executed.
* Util uses URLConnectionFactory as a static field and generates it only once.
So I think we can leave it as is.
So HDFS-14037.2.patch may work fine. What do you think about it?
was (Author: tasanuma0829):
[~brahmareddy] Sorry for late response.
* Considering that a new instance of RouterWebHdfsMethods is always created
every when a WebHDFS API is executed, storing the factory as an instance filed
may be no meaning.
* We can implement a method in DFSck to destroy the factory. But it may be
overkill because DFSck is a stand alone command, hadoop fsck, the lifetime is
short, and the factory is created only once when the command is executed.
* Util uses URLConnectionFactory as a static field and generates it only once.
So I think we can leave it as is.
So HDFS-14037.2.patch may be work fine. What do you think about it?
> Fix SSLFactory truststore reloader thread leak in URLConnectionFactory
> ----------------------------------------------------------------------
>
> Key: HDFS-14037
> URL: https://issues.apache.org/jira/browse/HDFS-14037
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: hdfs-client, webhdfs
> Reporter: Takanobu Asanuma
> Assignee: Takanobu Asanuma
> Priority: Major
> Attachments: HDFS-14037.1.patch, HDFS-14037.2.patch
>
>
> This is reported by [~yoshiata]. It is a similar issue as HADOOP-11368 and
> YARN-5309 in URLConnectionFactory.
> {quote}SSLFactory in newSslConnConfigurator and subsequently creates the
> ReloadingX509TrustManager instance which in turn starts a trust store
> reloader thread.
> However, the SSLFactory is never destroyed and hence the trust store reloader
> threads are not killed.
> {quote}
> We observed many leaked threads when we used swebhdfs via NiFi cluster.
> {noformat}
> "Truststore reloader thread" Id=221 TIMED_WAITING on null
> at java.lang.Thread.sleep(Native Method)
> at
> org.apache.hadoop.security.ssl.ReloadingX509TrustManager.run(ReloadingX509TrustManager.java:189)
> at java.lang.Thread.run(Thread.java:748)
> {noformat}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
