[
https://issues.apache.org/jira/browse/HBASE-10096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13845036#comment-13845036
]
Anoop Sam John commented on HBASE-10096:
----------------------------------------
Yes 96+ versions there is no problems. Also for the BC we will keep 0.94 as it
as. I will close this issue as Wont Fix. If you can guarentee the use of
changed version at client and server side then can appy the patch internally
and use. (The perf gain might be less only. Stil.... ) Pls keep the patch
attached to this jira so that some one wanted can use that.. Bacause of BC we
can not apply this to 0.94. Thanks for the finding and patch [~cuijianwei]
> make BulkDeleteResponse implement Writable to reduce the transfer size for rpc
> ------------------------------------------------------------------------------
>
> Key: HBASE-10096
> URL: https://issues.apache.org/jira/browse/HBASE-10096
> Project: HBase
> Issue Type: Improvement
> Components: Coprocessors, Performance
> Reporter: cuijianwei
> Attachments: HBASE-10096-0.94-v1.patch
>
>
> As [https://issues.apache.org/jira/browse/HBASE-6942] introduced,
> BulkDeleteEndpoint provides 'delete' method which performs like sql : "delete
> from table where ...". BulkDeleteEndpoint is efficient because it can
> complete scan and delete in one rpc and also could be implemented parallelly
> in different regions using coprocessor. BulkDeleteResponse is represents the
> result of BulkDeleteEndpoint.delete and will be serialized using a standard
> java serializable way. However, the serialized length of BulkDeleteResponse
> will be longer than one hundred byte length and may be not efficient enough
> to pass on the network. Therefore, is it better to make BulkDeleteResponse
> implement Writable interface and provide more efficient serialize method?
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
