hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Lars Hofhansl (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-12148) Remove TimeRangeTracker as point of contention when many threads writing a Store
Date Wed, 08 Oct 2014 03:16:34 GMT

    [ https://issues.apache.org/jira/browse/HBASE-12148?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14163004#comment-14163004
] 

Lars Hofhansl commented on HBASE-12148:
---------------------------------------

Yeah, we should either do AtomicLongs or synchronized + volatile. I was suggesting synchronized
+ volatile.
AtomicLong vs synchronized+volatile is interesting, we should test. AtomicLong (without synchronized)
is probably faster (CAS + volatile), especially when uncontended. In either case we'll eat
at least a memory barrier.

I would not trust code that compares anything between multiple threads that is not at least
volatile. Not sure of the memory model guarantees even timeline consistency between cores
without any for of memory barrier (synchronized, volatile, or Atomic*)


> Remove TimeRangeTracker as point of contention when many threads writing a Store
> --------------------------------------------------------------------------------
>
>                 Key: HBASE-12148
>                 URL: https://issues.apache.org/jira/browse/HBASE-12148
>             Project: HBase
>          Issue Type: Sub-task
>          Components: Performance
>    Affects Versions: 2.0.0, 0.99.1
>            Reporter: stack
>            Assignee: stack
>             Fix For: 2.0.0, 0.98.7, 0.99.1
>
>         Attachments: 12148.addendum.txt, 12148.txt, 12148.txt, 12148v2.txt, 12148v2.txt,
Screen Shot 2014-10-01 at 3.39.46 PM.png, Screen Shot 2014-10-01 at 3.41.07 PM.png
>
>




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

Mime
View raw message