hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Anoop Sam John (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-15562) Gc pause increases when the incoming requests are pooled
Date Fri, 01 Apr 2016 18:51:25 GMT

    [ https://issues.apache.org/jira/browse/HBASE-15562?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15222160#comment-15222160
] 

Anoop Sam John commented on HBASE-15562:
----------------------------------------

Some update here.
After some tuning with SurvivorRatio, we are able to get almost same avg GC pause time and
reduced #GC pauses, when the reqs read into BBs that we get from BBBpool..    So these are
off heap BBs.   We have extended PB to support offheap BB.   Also we have done an impl of
Offheap MSLAB.  Will give more test results soon.

> Gc pause increases when the incoming requests are pooled
> --------------------------------------------------------
>
>                 Key: HBASE-15562
>                 URL: https://issues.apache.org/jira/browse/HBASE-15562
>             Project: HBase
>          Issue Type: Sub-task
>            Reporter: ramkrishna.s.vasudevan
>
> This JIRA is to analyse why the GC pause time increases on introducing a BB pool on the
request side. This JIRA aims to track down the reason behind it. Having HBASE-15180 has an
increased effect in this GC pause time along with the Bb pool on the request side. 



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

Mime
View raw message