[
https://issues.apache.org/jira/browse/HBASE-10771?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13955379#comment-13955379
]
Nick Dimiduk commented on HBASE-10771:
--------------------------------------
bq. I will note that only you seemed to raise an objection to using the Netty 4
ByteBuf classes.
[~apurtell] I don't recall the specific comment to which you refer. My only
objection to Netty4 (or any other alternative, as they are) is a lack of
investigation. In no way did I expect such a comment to deter anyone from
conducting exactly that investigation!
The 3 alternatives I'm aware of are ByteBuffer (perhaps with reflection hacks I
experimented with a while back), Netty4 ByteBuf, and roll-our-own ByteRange. My
understanding of these options places ByteBuf at the top of my list, but I have
only anecdotal evidence to support its use. This is a fundamental component of
this system and I'd hope to see the decision driven by data proven through
investigation/evaluation as opposed to anecdotal evidence. My primary criteria
for a selection would be API compatibility down into HDFS and performance, and
really the former is driven by the latter.
All of this is just opinion of someone looking on. You three are doing the
actual good work here, so please continue doing it. Until such time as I (or
anyone else) am submitting patches along with you, please take my concerns as
they are: concerns voiced from the peanut-gallery to be considered only if
intelligent.
> Primitive type put/get APIs in ByteRange
> -----------------------------------------
>
> Key: HBASE-10771
> URL: https://issues.apache.org/jira/browse/HBASE-10771
> Project: HBase
> Issue Type: Improvement
> Reporter: Anoop Sam John
> Assignee: Anoop Sam John
> Fix For: 0.99.0
>
> Attachments: HBASE-10771.patch, HBASE-10771_V2.patch
>
>
> While doing HBASE-10713 I came across the need to write int/long (and read
> also) from a ByteRange. CellBlocks are backed by ByteRange. So we can add
> such APIs.
> Also as per HBASE-10750 we return a ByteRange from MSLAB and also discussion
> under HBASE-10191 suggest we can have BR backed HFileBlocks etc.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
