[
https://issues.apache.org/jira/browse/HDFS-3541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13397263#comment-13397263
]
Vinay commented on HDFS-3541:
-----------------------------
Hi Lee,
I agree your suggestion.
But as of now there is no suggestion to stop the PacketResponder.
Following things can happen with the proposed solution,
# initReplicaRecovery(..) call will interrupt the receiver thread, but it will
make the replica state to RUR and release the fsdataset lock.
# now PacketResponder may finalize the block, i.e. replica state will be
changed to FINALIZED,
# then updateReplicaUnderRecovery(..) call will fail because replica is not in
RUR state
I think we can restrict PacketResponder to finalize the block which is in RUR
by throwing exception. In this case updateReplicaUnderRecovery(..) will not
fail, and recovery will be success.
> Deadlock between recovery, xceiver and packet responder
> -------------------------------------------------------
>
> Key: HDFS-3541
> URL: https://issues.apache.org/jira/browse/HDFS-3541
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: data-node
> Affects Versions: 0.23.3, 2.0.1-alpha
> Reporter: suja s
> Assignee: Vinay
> Attachments: DN_dump.rar
>
>
> Block Recovery initiated while write in progress at Datanode side. Found a
> lock between recovery, xceiver and packet responder.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
