[
https://issues.apache.org/jira/browse/HDFS-6247?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14074047#comment-14074047
]
Vinayakumar B commented on HDFS-6247:
-------------------------------------
Hi [~clamb], Thanks for taking a look at the patch.
bq. I'm curious about why you are using a 5sec heartbeat interval. That seems
small relative to the timeout on the socket.
I thought it will be good enough to send the status. Since the total number of
block movements will be limited in balancing by bandwidth, I felt 5 second
interval will not going to add too much traffic.
How much you want to me increase? 30 sec would be fine?
> Avoid timeouts for replaceBlock() call by sending intermediate responses to
> Balancer
> ------------------------------------------------------------------------------------
>
> Key: HDFS-6247
> URL: https://issues.apache.org/jira/browse/HDFS-6247
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: balancer, datanode
> Affects Versions: 2.4.0
> Reporter: Vinayakumar B
> Assignee: Vinayakumar B
> Attachments: HDFS-6247.patch, HDFS-6247.patch, HDFS-6247.patch,
> HDFS-6247.patch
>
>
> Currently there is no response sent from target Datanode to Balancer for the
> replaceBlock() calls.
> Since the Block movement for balancing is throttled, complete block movement
> will take time and this could result in timeout at Balancer, which will be
> trying to read the status message.
>
> To Avoid this during replaceBlock() call in in progress Datanode can send
> IN_PROGRESS status messages to Balancer to avoid timeouts and treat
> BlockMovement as failed.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
