hadoop-hdfs-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lei (Eddy) Xu (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (HDFS-13039) StripedBlockReader#createBlockReader leaks socket on IOException
Date Fri, 19 Jan 2018 18:59:01 GMT

     [ https://issues.apache.org/jira/browse/HDFS-13039?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Lei (Eddy) Xu updated HDFS-13039:
---------------------------------
    Status: Patch Available  (was: Open)

Upload the fix to close {{Peer}} on {{IOException}}.

[~manojg] [~Sammi] could you help to review it?

> StripedBlockReader#createBlockReader leaks socket on IOException
> ----------------------------------------------------------------
>
>                 Key: HDFS-13039
>                 URL: https://issues.apache.org/jira/browse/HDFS-13039
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: datanode, erasure-coding
>    Affects Versions: 3.0.0
>            Reporter: Lei (Eddy) Xu
>            Assignee: Lei (Eddy) Xu
>            Priority: Critical
>         Attachments: HDFS-13039.00.patch
>
>
> When running EC on one cluster, DataNode has millions of {{CLOSE_WAIT}} connections
> {code:java}
> $ grep CLOSE_WAIT lsof.out | wc -l
> 10358700
> // All CLOSW_WAITs belong to the same DataNode process (pid=88527)
> $ grep CLOSE_WAIT lsof.out | awk '{print $2}' | sort | uniq
> 88527
> {code}
> And DN can not open any file / socket, as shown in the log:
> {noformat}
> 2018-01-19 06:47:09,424 WARN io.netty.channel.DefaultChannelPipeline: An exceptionCaught()
event was fired, and it reached at the tail of the pipeline. It usually means the last handler
in the pipeline did not handle the exception.
> java.io.IOException: Too many open files
>         at sun.nio.ch.ServerSocketChannelImpl.accept0(Native Method)
>         at sun.nio.ch.ServerSocketChannelImpl.accept(ServerSocketChannelImpl.java:422)
>         at sun.nio.ch.ServerSocketChannelImpl.accept(ServerSocketChannelImpl.java:250)
>         at io.netty.channel.socket.nio.NioServerSocketChannel.doReadMessages(NioServerSocketChannel.java:135)
>         at io.netty.channel.nio.AbstractNioMessageChannel$NioMessageUnsafe.read(AbstractNioMessageChannel.java:75)
>         at io.netty.channel.nio.NioEventLoop.processSelectedKey(NioEventLoop.java:563)
>         at io.netty.channel.nio.NioEventLoop.processSelectedKeysOptimized(NioEventLoop.java:504)
>         at io.netty.channel.nio.NioEventLoop.processSelectedKeys(NioEventLoop.java:418)
>         at io.netty.channel.nio.NioEventLoop.run(NioEventLoop.java:390)
>         at io.netty.util.concurrent.SingleThreadEventExecutor$5.run(SingleThreadEventExecutor.java:742)
>         at io.netty.util.concurrent.DefaultThreadFactory$DefaultRunnableDecorator.run(DefaultThreadFactory.java:145)
>         at java.lang.Thread.run(Thread.java:748)
> {noformat}



--
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


Mime
View raw message