[
https://issues.apache.org/jira/browse/HDFS-9588?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15088412#comment-15088412
]
Arpit Agarwal commented on HDFS-9588:
-------------------------------------
Hi [~anu], my comments below.
# Missing check for presence of {{maxDiskBandwidth}} when deserializing the
protobuf message.
# Instead of enumerating the error codes in
{{SubmitDiskBalancerPlanResponseProto}} can we just have the DataNode throw
appropriate exceptions, when the server-side implementation is done?
# Missing Javadoc for {{ClientDatanodeProtocol#submitDiskBalancerPlan}}.
Perhaps also document any uniqueness constraints for planID here.
# Nitpick: Should {{DataNode#submitDiskBalancerPlan}} throw
{{NotImplementedException}} until it is implemented, instead of returning
{{OK}}?
# Is the plan version used currently?
Stylistic point - we can make most of the protobuf fields {{optional}}?
https://developers.google.com/protocol-buffers/docs/proto?hl=en#specifying-field-rules.
It does require more code to handle {{optional}} fields when deserializing the
message. Your call, I don't think we have any project-wide guideline.
> DiskBalancer : Add submitDiskbalancer RPC
> -----------------------------------------
>
> Key: HDFS-9588
> URL: https://issues.apache.org/jira/browse/HDFS-9588
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: balancer & mover
> Affects Versions: HDFS-1312
> Reporter: Anu Engineer
> Assignee: Anu Engineer
> Attachments: HDFS-9588-HDFS-1312.001.patch,
> HDFS-9588-HDFS-1312.002.patch
>
>
> Add a data node RPC that allows client to submit a diskbalancer plan to data
> node.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
