Return-Path: X-Original-To: apmail-hbase-issues-archive@www.apache.org Delivered-To: apmail-hbase-issues-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 7E819D402 for ; Fri, 7 Sep 2012 21:42:10 +0000 (UTC) Received: (qmail 97156 invoked by uid 500); 7 Sep 2012 21:42:09 -0000 Delivered-To: apmail-hbase-issues-archive@hbase.apache.org Received: (qmail 97104 invoked by uid 500); 7 Sep 2012 21:42:09 -0000 Mailing-List: contact issues-help@hbase.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Delivered-To: mailing list issues@hbase.apache.org Received: (qmail 97057 invoked by uid 99); 7 Sep 2012 21:42:08 -0000 Received: from arcas.apache.org (HELO arcas.apache.org) (140.211.11.28) by apache.org (qpsmtpd/0.29) with ESMTP; Fri, 07 Sep 2012 21:42:08 +0000 Date: Sat, 8 Sep 2012 08:42:08 +1100 (NCT) From: "stack (JIRA)" To: issues@hbase.apache.org Message-ID: <945350096.52487.1347054128966.JavaMail.jiratomcat@arcas> In-Reply-To: <338363458.52051.1347049447585.JavaMail.jiratomcat@arcas> Subject: [jira] [Commented] (HBASE-6744) Per table balancing could cause regions unbalanced overall MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit X-JIRA-FingerPrint: 30527f35849b9dde25b450d4833f0394 [ https://issues.apache.org/jira/browse/HBASE-6744?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13451016#comment-13451016 ] stack commented on HBASE-6744: ------------------------------ +1 on making SLB default now. Testing for 0.96.0 should shake out any bugs. On its face alone, it has a potential far in excess of the current, hard to untangle, served-us well up-to-this, but needs-to-be retired now current balancer. > Per table balancing could cause regions unbalanced overall > ---------------------------------------------------------- > > Key: HBASE-6744 > URL: https://issues.apache.org/jira/browse/HBASE-6744 > Project: HBase > Issue Type: Improvement > Reporter: Jimmy Xiang > > Per table balancing just balances regions based on tables. However, overall, regions could be seriously unbalanced. > For example, if you shutdown all most all region serves in a cluster, then create tons of new tables (no region pre-split), then start up all region servers. You will see the regions won't move to other region servers since they are balanced per table (only one region for a table at this moment). > If we can make the balance algorithm sophisticated enough, we don't need the configuration hbase.master.loadbalance.bytable. We can do the regular and bytable balancing at the same time. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira