hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nicolas Spiegelberg (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-3620) Make HBCK Faster
Date Sat, 12 Mar 2011 18:21:59 GMT

    [ https://issues.apache.org/jira/browse/HBASE-3620?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13006071#comment-13006071

Nicolas Spiegelberg commented on HBASE-3620:

@Ted: I'm not sure about what problem this causes.  The standard Java API allows you to create
Thread Pools using this method & java.util.concurrent.ExecutorServer.Executors.  Most
of the codebase uses Executors.newFixedThreadPool(), which is a factory wrapper around the
TPE API.  In this case, you don't need a fixed threadcount because you might have a small
cluster or large cluster.  You want your TPE to dynamic grow for your scale.

If you're interested in where we use Thread Pools, a better search is for 'ExecutorService'

> Make HBCK Faster
> ----------------
>                 Key: HBASE-3620
>                 URL: https://issues.apache.org/jira/browse/HBASE-3620
>             Project: HBase
>          Issue Type: Improvement
>            Reporter: Nicolas Spiegelberg
>            Assignee: Nicolas Spiegelberg
>            Priority: Minor
>             Fix For: 0.90.2, 0.92.0
>         Attachments: hbck_perf.patch
> Make the HBCK utility contact all region servers & HDFS directories in parallel.
This will speedup hbck processing, especially when there are lots of region servers.

This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message