Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Date: Fri, 27 Mar 2015 04:53:52 +0000 (UTC)
From: "Lars Hofhansl (JIRA)" <jira@apache.org>
To: issues@hbase.apache.org
Message-ID: <JIRA.12783345.1426793969000.46406.1427432032766@Atlassian.JIRA>
In-Reply-To: <JIRA.12783345.1426793969000@Atlassian.JIRA>
References: <JIRA.12783345.1426793969000@Atlassian.JIRA>
 <JIRA.12783345.1426793969503@arcas>
Subject: [jira] [Commented] (HBASE-13291) Lift the scan ceiling
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HBASE-13291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14383331#comment-14383331 ] 

Lars Hofhansl commented on HBASE-13291:
---------------------------------------

bq.  That work seems to address other issues than what I am seeing here.
Yeah. Was thinking we use the index I build there to avoid the repeated getKeyValueLen calls.

I'm surprised that that StoreScanner.next() is only in there with 6% and StoreScanner.peek() not all. By any chance is this still with the lock coarsening patch? (or you're testing with kinda larger column values too, so maybe that's why).

> Lift the scan ceiling
> ---------------------
>
>                 Key: HBASE-13291
>                 URL: https://issues.apache.org/jira/browse/HBASE-13291
>             Project: HBase
>          Issue Type: Improvement
>          Components: Scanners
>    Affects Versions: 1.0.0
>            Reporter: stack
>            Assignee: stack
>         Attachments: 13291.inlining.txt, Screen Shot 2015-03-26 at 12.12.13 PM.png, Screen Shot 2015-03-26 at 3.39.33 PM.png, hack_to_bypass_bb.txt, nonBBposAndInineMvccVint.txt, q (1).png, traces.7.svg, traces.filterall.svg, traces.nofilter.svg, traces.small2.svg, traces.smaller.svg
>
>
> Scanning medium sized rows with multiple concurrent scanners exhibits interesting 'ceiling' properties. A server runs at about 6.7k ops a second using 450% of possible 1600% of CPUs  when 4 clients each with 10 threads doing scan 1000 rows.  If I add '--filterAll' argument (do not return results), then we run at 1450% of possible 1600% possible but we do 8k ops a second.
> Let me attach flame graphs for two cases. Unfortunately, there is some frustrating dark art going on. Let me try figure it... Filing issue in meantime to keep score in.


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