You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "stack (JIRA)" <ji...@apache.org> on 2015/09/25 18:02:04 UTC
[jira] [Commented] (HBASE-14490) [RpcServer] reuse buffer
[ https://issues.apache.org/jira/browse/HBASE-14490?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14908238#comment-14908238 ]
stack commented on HBASE-14490:
-------------------------------
bq. ByteBuffer.array might contain garbage. Also keeping objects tend to require Full GC to collect the objects.
Isn't that ok as long as the length-of-read accounting is properly kept up [~ikeda]? (There is a 'clear' on the end of the patch if the buffer is not too large; i.e. we do not retain very large buffers).
[~gzh1992n] The failures above seem related. Is it correct that you flip the dataLengthBuffer each time through?
I think this could make for a nice optimization [~gzh1992n] and a good place to dig in. Should we swap in an instance of BoundedByteBufferPool here? (Could make for a nice difference especially with say the nice finding by [~anoop.hbase] over in HBASE-13819)
> [RpcServer] reuse buffer
> ------------------------
>
> Key: HBASE-14490
> URL: https://issues.apache.org/jira/browse/HBASE-14490
> Project: HBase
> Issue Type: Improvement
> Components: IPC/RPC
> Affects Versions: 2.0.0
> Reporter: Zephyr Guo
> Assignee: Zephyr Guo
> Priority: Minor
> Attachments: HBASE-14490-v1.patch
>
>
> Reuse buffer to read request.It's not necessary free data's buffer for each RPC.It maybe can cause memory fragmentation.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)