[
https://issues.apache.org/jira/browse/HDFS-7116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14555625#comment-14555625
]
Rakesh R commented on HDFS-7116:
--------------------------------
*Initial design thoughts:*
Case-1) Datanode will inform Namenode about the
{{dfs.datanode.balance.bandwidthPerSec}} configured value through the
heartbeat. Initially I thought of updating Namenode only once during the
DatanodeRegistration process, but on a second thought I decided to send it
through heartbeat for consistency. On receiving the heartbeat, Namenode
HeartbeatManager will update the bandwidth value in {{DatanodeInfo#bandwidth }}
field. Then continue to the existing logic of checking a new bandwidth value
from admin. Here, this check will be like,
{code}
if ( DatanodeInfo#newBalancerBandwidth > 0 && DatanodeInfo#newBalancerBandwidth
!= DatanodeInfo#bandwidth )
{code}
then sent UPDATE_BANDWIDTH command to Datanode.
Case-2) Administrator sends {{-setBalancerBandwidth}} command. Now, this
information will be keeping in a new field {{newBalancerBandwidth}} in the
{{DatanodeInfo}} object.
Could you please review the algorithm and appreciate review comments. Thanks!
> Add a command to get the bandwidth of balancer
> ----------------------------------------------
>
> Key: HDFS-7116
> URL: https://issues.apache.org/jira/browse/HDFS-7116
> Project: Hadoop HDFS
> Issue Type: New Feature
> Components: balancer & mover
> Reporter: Akira AJISAKA
> Assignee: Rakesh R
>
> Now reading logs is the only way to check how the balancer bandwidth is set.
> It would be useful for administrators if they can get the parameter via CLI.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
