accumulo-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Josh Elser (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (ACCUMULO-4458) lock contention around configuration settings impacts tablet server performance
Date Thu, 15 Sep 2016 00:23:20 GMT

    [ https://issues.apache.org/jira/browse/ACCUMULO-4458?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15491870#comment-15491870
] 

Josh Elser commented on ACCUMULO-4458:
--------------------------------------

bq. Right now I do extract all of the properties that are set

Ahh, apologies. I see this now. That leads me to wonder...

{code}
-    String value = getXmlConfig().get(key);
+    /* Check the available-on-load configs and fall-back to the possibly-update Configuration
object. */
+    String value = staticConfigs.containsKey(key) ? staticConfigs.get(key) : getXmlConfig().get(key);
{code}

Why can't this just be {{String value = staticConfigs.get(key)}} then? We know that we have
an explicit Accumulo config key already (given the method signature). Am I missing something
else?

bq. yeah, I was confused for awhile about what was going on with zooCache, since it was static
but was initialized in the constructor of the singleton. Eventually decided it just must have
been a mistake.

Makes sense, the bit-rot is real ;). HdfsZooInstance looks like it was always like this (from
initial import). 

> lock contention around configuration settings impacts tablet server performance
> -------------------------------------------------------------------------------
>
>                 Key: ACCUMULO-4458
>                 URL: https://issues.apache.org/jira/browse/ACCUMULO-4458
>             Project: Accumulo
>          Issue Type: Bug
>          Components: tserver
>    Affects Versions: 1.7.1, 1.8.0
>            Reporter: Sean Busbey
>            Assignee: Sean Busbey
>            Priority: Critical
>             Fix For: 1.7.3, 1.8.1, 2.0.0
>
>         Attachments: ACCUMULO-4458-1.7.v1.patch
>
>
> While investigating a pretty severe performance regression comparing YCSB against 1.6
and 1.7, I found a fair bit of lock contention around getting configuration values. This was
improved by ACCUMULO-4388, but various threads eventually all started contending on the configuration
values expected from the site xml files.



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

Mime
View raw message