You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-user@hadoop.apache.org by sunww <sp...@outlook.com> on 2014/07/04 04:31:31 UTC

about rack awareness

HiI have two questions about rack awareness.
1、I installed a hadoop2.2 , hbase0.96 cluster, which have 2 racks. rack1 have 10 machines , rack2 have 5 machines. Because of 3 replicas, rack2 will store more data then rack1.Maybe I need to move some machines from  rack2 to rack1?If I had more racks, only one have 5 machines and the others have 10 machines, Will data unbalance more serious? 
2、If some racks share the same  power and switch, is rack awareness necessary ?

Thanks  		 	   		  

Re: about rack awareness

Posted by Kai Voigt <k...@123.org>.
Rack Awareness actually should be called Switch Awareness, and that’s what people typically do: Nodes in a rack are at the same switch, also you should have balanced capacity across racks/switches for a balanced distribution.

So you definitely want to rearrange your setup, especially if your cluster grows larger.

Kai

Am 04.07.2014 um 03:31 schrieb sunww <sp...@outlook.com>:

> Hi
> I have two questions about rack awareness.
> 
> 1、I installed a hadoop2.2 , hbase0.96 cluster, which have 2 racks. 
> rack1 have 10 machines , rack2 have 5 machines. Because of 3 replicas, rack2 will store more data then rack1.
> Maybe I need to move some machines from  rack2 to rack1?
> If I had more racks, only one have 5 machines and the others have 10 machines, Will data unbalance more serious? 
> 
> 2、If some racks share the same  power and switch, is rack awareness necessary ?
> 
> 
> Thanks 

Kai Voigt			Am Germaniahafen 1			k@123.org
					24143 Kiel					+49 160 96683050
					Germany						@KaiVoigt


Re: about rack awareness

Posted by Kai Voigt <k...@123.org>.
Rack Awareness actually should be called Switch Awareness, and that’s what people typically do: Nodes in a rack are at the same switch, also you should have balanced capacity across racks/switches for a balanced distribution.

So you definitely want to rearrange your setup, especially if your cluster grows larger.

Kai

Am 04.07.2014 um 03:31 schrieb sunww <sp...@outlook.com>:

> Hi
> I have two questions about rack awareness.
> 
> 1、I installed a hadoop2.2 , hbase0.96 cluster, which have 2 racks. 
> rack1 have 10 machines , rack2 have 5 machines. Because of 3 replicas, rack2 will store more data then rack1.
> Maybe I need to move some machines from  rack2 to rack1?
> If I had more racks, only one have 5 machines and the others have 10 machines, Will data unbalance more serious? 
> 
> 2、If some racks share the same  power and switch, is rack awareness necessary ?
> 
> 
> Thanks 

Kai Voigt			Am Germaniahafen 1			k@123.org
					24143 Kiel					+49 160 96683050
					Germany						@KaiVoigt


Re: about rack awareness

Posted by Kai Voigt <k...@123.org>.
Rack Awareness actually should be called Switch Awareness, and that’s what people typically do: Nodes in a rack are at the same switch, also you should have balanced capacity across racks/switches for a balanced distribution.

So you definitely want to rearrange your setup, especially if your cluster grows larger.

Kai

Am 04.07.2014 um 03:31 schrieb sunww <sp...@outlook.com>:

> Hi
> I have two questions about rack awareness.
> 
> 1、I installed a hadoop2.2 , hbase0.96 cluster, which have 2 racks. 
> rack1 have 10 machines , rack2 have 5 machines. Because of 3 replicas, rack2 will store more data then rack1.
> Maybe I need to move some machines from  rack2 to rack1?
> If I had more racks, only one have 5 machines and the others have 10 machines, Will data unbalance more serious? 
> 
> 2、If some racks share the same  power and switch, is rack awareness necessary ?
> 
> 
> Thanks 

Kai Voigt			Am Germaniahafen 1			k@123.org
					24143 Kiel					+49 160 96683050
					Germany						@KaiVoigt


Re: about rack awareness

Posted by Kai Voigt <k...@123.org>.
Rack Awareness actually should be called Switch Awareness, and that’s what people typically do: Nodes in a rack are at the same switch, also you should have balanced capacity across racks/switches for a balanced distribution.

So you definitely want to rearrange your setup, especially if your cluster grows larger.

Kai

Am 04.07.2014 um 03:31 schrieb sunww <sp...@outlook.com>:

> Hi
> I have two questions about rack awareness.
> 
> 1、I installed a hadoop2.2 , hbase0.96 cluster, which have 2 racks. 
> rack1 have 10 machines , rack2 have 5 machines. Because of 3 replicas, rack2 will store more data then rack1.
> Maybe I need to move some machines from  rack2 to rack1?
> If I had more racks, only one have 5 machines and the others have 10 machines, Will data unbalance more serious? 
> 
> 2、If some racks share the same  power and switch, is rack awareness necessary ?
> 
> 
> Thanks 

Kai Voigt			Am Germaniahafen 1			k@123.org
					24143 Kiel					+49 160 96683050
					Germany						@KaiVoigt