[Typo3-dev] Clustering

Mathias Schreiber [wmdb] mathias.schreiber at wmdb.de
Wed Aug 3 17:35:38 CEST 2005


Sven Wilhelm wrote:
> for a database server this can make sense, but it is not the golden way
> to go. If your system is designed to scale and you have a controlled way
> for changes extending your system by adding a new node make more sense
> and scales better.

If adding a node works.
This is a sensitive subject because noone has real information on how 
typo3 works in a high load environment with mysql cluster running.
We once had the problem to face more than 500.000 PIs per hour with 
typo3 and went for IBM blades.
They keep track of rebuilding the data and keeping the data in sync all 
by themselves (funny, just like they say in their marketing paper) and 
this works fine.

I know that jhh once had tried some clustering solution from a 3rd party 
  company but ran into trouble because this solution wasn't able to sync 
data bigger than 64kb.

-- 
if ($GLOBALS['TSFE']->fe_user->user['ahnung'] == 0) {
	$this->fresseHalten = 1;
}




More information about the TYPO3-dev mailing list