[
https://issues.apache.org/jira/browse/HDFS-6569?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14044680#comment-14044680
]
Kihwal Lee commented on HDFS-6569:
----------------------------------
{{TestClientProtocolForPipelineRecovery#testPipelineRecoveryOnOOB}} is supposed
to test the case, but in this test case the thread will be in epoll_wait(), not
read(). I guess that's the difference.
bq. One possible fix is to send OOB message before interrupting DataXceiver
threads.
I explored this approach before, I will see what I can do.
> OOB message can't be sent to the client when DataNode shuts down for upgrade
> ----------------------------------------------------------------------------
>
> Key: HDFS-6569
> URL: https://issues.apache.org/jira/browse/HDFS-6569
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: datanode
> Affects Versions: 3.0.0, 2.4.0
> Reporter: Brandon Li
>
> The socket is closed too early before the OOB message can be sent to client,
> which causes the write pipeline failure.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
