hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nicolas Liochon (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-10449) Wrong execution pool configuration in HConnectionManager
Date Tue, 04 Feb 2014 09:32:10 GMT

    [ https://issues.apache.org/jira/browse/HBASE-10449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13890522#comment-13890522
] 

Nicolas Liochon commented on HBASE-10449:
-----------------------------------------

Note that we're not exactly changing a default. 
The code wanted to do:
 {panel}Create up to 'max' (default: 256) threads. Expires them if they are not used for 10
seconds, excepted for 'core' (default 0) of them. If there is more than 'max' tasks, queue
them.{panel} 

Actually it was doing:
{panel}Create a single thread, queue all the tasks for this thread.{panel}

So the patch actually implements that was supposed to be implemented (or tries to implement
it at least :-) ). I

Moreover, it's a regression from HBASE-9917, so actually 96.0 really uses 256 threads. It's
a *96.1* issue only. But yes, it does have an impact on performances, and this impact can
be good or bad. That's why I would like it to be in the .98 RC, and also why I think it's
simpler to have the same defaults on all versions.

Lastly, and unrelated, we didn't have a limit of the number of threads before the .96. I'm
wondering if we don't have an impact if a server hangs. The client may ends up with all its
connections stuck to this server, until it timeouts.

> Wrong execution pool configuration in HConnectionManager
> --------------------------------------------------------
>
>                 Key: HBASE-10449
>                 URL: https://issues.apache.org/jira/browse/HBASE-10449
>             Project: HBase
>          Issue Type: Bug
>          Components: Client
>    Affects Versions: 0.98.0, 0.99.0, 0.96.1.1
>            Reporter: Nicolas Liochon
>            Assignee: Nicolas Liochon
>            Priority: Critical
>             Fix For: 0.98.0, 0.96.2, 0.99.0
>
>         Attachments: HBASE-10449.v1.patch
>
>
> There is a confusion in the configuration of the pool. The attached patch fixes this.
This may change the client performances, as we were using a single thread.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

Mime
View raw message