hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ramkrishna.s.vasudevan (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-16703) Explore object pooling of SeekerState
Date Mon, 26 Sep 2016 05:39:20 GMT

    [ https://issues.apache.org/jira/browse/HBASE-16703?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15522125#comment-15522125
] 

ramkrishna.s.vasudevan commented on HBASE-16703:
------------------------------------------------

[~apurtell]
I was working on this area long time back regarding the byte[] that was getting recreated
if the size grows above INITIAL_BUFFER_SIZE. But that was to see if we can avoid the recreation.
But I think we can atleast avoid the recreation of the buffer because per SeekerState we are
going to keep filling this fixed byte array as per the cell that is retrieved.
Let me know if I can take this up as this is unassigned. If you are planning to work on it,
pls feel free to do so.

> Explore object pooling of SeekerState
> -------------------------------------
>
>                 Key: HBASE-16703
>                 URL: https://issues.apache.org/jira/browse/HBASE-16703
>             Project: HBase
>          Issue Type: Task
>            Reporter: Andrew Purtell
>
> In read workloads 35% of the allocation pressure produced by servicing RPC requests,
when block encoding is enabled, comes from BufferedDataBlockEncoder$SeekerState.<init>,
where we allocate two byte arrays of INITIAL_KEY_BUFFER_SIZE in length. There's an opportunity
for object pooling of SeekerState here. Subsequent code checks if those byte arrays are sized
sufficiently to handle incoming data to copy. The arrays will be resized if needed. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message