Mailing-List: contact user-help@hbase.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@hbase.apache.org
Received-SPF: error (athena.apache.org: local policy)
MIME-Version: 1.0
In-Reply-To: <17B77116-A83E-449D-84AF-F8FDE88B04E2@gmail.com>
References: 
 <CACeqxwSp5HeauMkZuqKeAY1kZ2bZ+ye86VBKKgC_u0H53jTuFg@mail.gmail.com>
 <167F3C3F-8013-4258-B3D6-59418F8FCFDB@gmail.com>
 <CALr1C9pduh-ygBTTCM8gUt8ueEC8kgObDWVQ46CGSOQOO3n5Hg@mail.gmail.com>
 <CACeqxwRxShQ5SnONrTs+Bjs1UdRrqSxDgDFhoFa2X_i=0khb6w@mail.gmail.com>
 <17B77116-A83E-449D-84AF-F8FDE88B04E2@gmail.com>
From: Jean-Marc Spaggiari <jean-marc@spaggiari.org>
Date: Thu, 4 Jul 2013 09:16:14 -0400
Message-ID: 
 <CAPQV63Vgnivovh_rYayERFV4azibqq31QJm5Kyk1gmjiwo+HNQ@mail.gmail.com>
Subject: Re: unbalance number fo region in cluster
To: user@hbase.apache.org
Content-Type: text/plain; charset=UTF-8

Can you please also checked on the HBase Web UI?
Do you have regions in transition?

Maybe the balancer already gave te balance plan, and regions are
moving, slowly, to their destinations? How big are those regions? I
mean, size in MB/GB and number of rows?


JM

2013/7/4 Ted Yu <yuzhihong@gmail.com>:
> Can you check load balancer related log lines in master log and put them on pastebin so that root cause can be diagnosed ?
>
> Which HBase version are you using ?
>
> Thanks
>
> On Jul 3, 2013, at 11:44 PM, ch huang <justlooks@gmail.com> wrote:
>
>> the load balancer process as i see is very slow i have a big table  ,it's
>> region distrabute is  node1: 5 ,node2: 22 ,node3: 5 ,after 4 hours ,it's
>> became
>> node1: 6  ,node2: 20 ,node1: 6, change is very very slow!
>>
>> On Thu, Jul 4, 2013 at 1:54 PM, Azuryy Yu <azuryyyu@gmail.com> wrote:
>>
>>> I supposed load balancer is enabled and there is no pre-split.
>>>
>>> LB run periodicly, (5 mininutes by default), so after insert bulk data into
>>> HBase, the table is unbalance but there will be balanced eventually after
>>> round-robin load balancer running.
>>>
>>>
>>> On Thu, Jul 4, 2013 at 1:25 PM, Ted Yu <yuzhihong@gmail.com> wrote:
>>>
>>>> Did you presplit your table ?
>>>>
>>>> Was load balancer enabled ?
>>>>
>>>> What HBase version do you use ?
>>>>
>>>> Thanks
>>>>
>>>> On Jul 3, 2013, at 10:21 PM, ch huang <justlooks@gmail.com> wrote:
>>>>
>>>>> i insert 2GB data into my hbase cluster ,but i find the region is not
>>>>> balance,is insert operation always cause the region number different on
>>>>> region server?
>>>>>
>>>>> here is my 60010 output
>>>>>
>>>>> Table Regions  Name Region Server Start Key End Key
>>>>> demo2,,1372914631698.f42281ec44d113df3e644d5ecdb35366.
>>>>> CH36:60030<http://ch36:60030/>
>>>>> rows23703
>>> demo2,rows23703,1372914684582.c3373791daec95b849ba2908c517e045.
>>>>> CH35:60030 <http://ch35:60030/> rows23703 rows347161
>>>>> demo2,rows347161,1372914714248.dd9a6c39f65ad4a879638e1bfec76016.
>>>>> CH34:60030<http://ch34:60030/>
>>>>> rows347161 rows428378
>>>>> demo2,rows428378,1372914745946.41a4669f36bc99ef1592b2df13c0a948.
>>>>> CH36:60030<http://ch36:60030/>
>>>>> rows428378 rows512589
>>>>> demo2,rows512589,1372914776151.a4363bdd7403633c5775931ab32a7da8.
>>>>> CH36:60030<http://ch36:60030/>
>>>>> rows512589 rows598644
>>>>> demo2,rows598644,1372914806469.729f52018aac909e8080f00b8d698e2d.
>>>>> CH36:60030<http://ch36:60030/>
>>>>> rows598644 rows684871
>>>>> demo2,rows684871,1372914836812.d9cb1760fe3302500034913ff813e0f8.
>>>>> CH35:60030<http://ch35:60030/>
>>>>> rows684871 rows771559
>>>>> demo2,rows771559,1372914836812.ac16ff1404cff5c0db3d2e3b0661ffcd.
>>>>> CH36:60030<http://ch36:60030/>
>>>>> rows771559 Regions by Region Server  Region Server Region Count
>>>>> http://CH35:60030/ <http://ch35:60030/> 2 http://CH36:60030/<http://ch36:60030/>5
>>> http://CH34:60030/ <http://ch34:60030/> 1
>>>