MySQL Forums
Forum List  »  NDB clusters

data repartition in cluster
Posted by: Pierre COLOMBIER
Date: June 01, 2005 03:49AM

I'm making some tests with different numbers of ndb nodes.

for the moment, (as i'm only doing tests) ther are no replicas


i've got , lets's say 750 Mo of data.

if i use 1 node

node 1 : 800Mo

if i use 2 nodes

node 1 : 400Mo
node 2 : 400Mo

if I use 3 nodes

node 1 : 200Mo
node 2 : 400Mo <<???
node 3 : 200Mo

If i use 4 nodes

node 1 : 200Mo
node 2 : 200Mo
node 3 : 200Mo
node 4 : 200Mo

If I use 6 nodes

node 1 : 100Mo
node 2 : 100Mo
node 3 : 200Mo
node 4 : 200Mo
node 5 : 100Mo
node 6 : 100Mo

If I use 8 nodes

node 1 : 120Mo
node 2 : 120Mo
node 3 : 120Mo
node 4 : 120Mo
node 5 : 280Ko <<<??? yes Kilo not Mega !
node 6 : 120Mo
node 7 : 120Mo
node 8 : 120Mo


I don't really understand how the data is placed on different nodes. isn't it equal repartition ? Are theses differences normal ?
Is it necessary to use an even number of nodegroups (or a power of 2) ?

Options: ReplyQuote


Subject
Views
Written By
Posted
data repartition in cluster
1551
June 01, 2005 03:49AM
1037
October 30, 2006 02:38AM


Sorry, you can't reply to this topic. It has been closed.

Content reproduced on this site is the property of the respective copyright holders. It is not reviewed in advance by Oracle and does not necessarily represent the opinion of Oracle or any other party.