[
https://issues.apache.org/jira/browse/ACCUMULO-2837?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14480872#comment-14480872
]
Josh Elser commented on ACCUMULO-2837:
--------------------------------------
Increased priority since it's a nice performance impact.
> Scanning empty table can be optimized
> -------------------------------------
>
> Key: ACCUMULO-2837
> URL: https://issues.apache.org/jira/browse/ACCUMULO-2837
> Project: Accumulo
> Issue Type: Bug
> Affects Versions: 1.6.0
> Reporter: Keith Turner
> Assignee: Keith Turner
> Priority: Critical
> Labels: performance
> Fix For: 1.6.3, 1.8.0, 1.7.1
>
>
> I wrote an application that repeatedly scanned random rows in an empty table
> at one point. It did not know the table was empty. I was jstacking the
> tserver and repeatedly saw stack traces in a few places in the code.
> * the audit canScan method
> * DefaultConfiguration.get
> * Kept seeing Password token ungziping the password??????
> After fixing all of these issues (except for system token) I ran a simple
> test I saw times go from ~1.7s to ~1.3s for a simple test I wrote to recreate
> this issue.
> I am not sure how much of an impact the gziping the password has. I was just
> sampling the stack traces manually. I seems like I saw the password token
> gzip stack traces less frequently, but I did catch it multiple jstack calls.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
