Home Page for the TeradataForum
 

Archives of the TeradataForum

Message Posted: Tue, 04 Jun 2002 @ 14:46:42 GMT


     
  <Prev Next>   <<First <Prev
Next>
Last>>
 


Subj:   Re: Optimal node configuration
 
From:   Charles Farley

Hello-

Having lived in an environment that had "incomplete" cliques on a production system, I will say this, configuring one clique as a "full" (four nodes), will have you seeing performance loss as though you had lost a node of the other clique (in the 4-4-2, or other with a "full" clique). We never realized how this hindered the system until we had a system with only full cliques on it, the difference was noticeable. I guess the only way I can really put it is how the NCR rep put it to me, "you're only as fast as your slowest clique..." (this really applies in systems where a node is down). We now have a policy of purchasing upgrades of only full cliques.

I may be way off base with this, this is my own experience, and I am more just a lowly sys admin than DBA, but the clique means a lot to the system and how it handles I/O and other functions that are related to having the complete clique. Much is dependent on the relationship between the nodes in a clique over and above the function of the database as an application on the running system.

In the end, get all the horsepower you can when you can, but I think you may see some hinderance of the system if you use incomplete cliques, especially for upgrades.


Hope this makes sense, I'm rather tired, thank you for your time and attention,

loadc



     
  <Prev Next>   <<First <Prev
Next>
Last>>
 
 
 
 
 
 
 
 
 
  
  Top Home Privacy Feedback  
 
 
Copyright for the TeradataForum (TDATA-L), Manta BlueSky    
Copyright 2016 - All Rights Reserved    
Last Modified: 15 Jun 2023