Hi,
My team and I have developed a nifty little app, which somewhat unfortunately, requires some time to process the data (15 million rows x 5 million rows x 3 million x 157 rows) when joining the queries.
We have worked very hard and dedicated trying to optimize the queries as good as we can and I really feel we have made enormous progress there. The problem is that the extent of the data growth is quite large and even though we make progress the data growth eats that progress slowly.
The setup looks as this:
2 webservers load-balancing the load from the users from a webgui (not that quite many users running the "monsterqueries" above)
2 App Servers (constantly adding new data from data-source)
2 DB-servers which at this moment runs MySQL 4.0 and does replication between them. (The hardware is 2 x Xeon 3,06GHZ 2GB RAM and RAIDED disks)
My questions are these;
1 ) Would I be better of with MySQL cluster than my replication set? I've read the whitepapers of it and it seem impressing. Does anyone have any experience of MySQL cluster? Is it overkill for us to use?
2) Is it true that it is embedded in the 4.1 Alpha? I tried to look for prices when using mysql cluster but found none, am I correct when assuming that it is the regular price even for MySQL cluster use?
3 ) Has anyone tried a setup? Is it hard and time consuming?
4 ) From what I've read on Mysqls homepage it seems like subqueries is a new function starting from 4.1. Are there any speed differences between joins (inner, left, right, outer) and subqueries?
5 ) When doing some queries against our database it takes more or less 2-3 seconds before results. We have tested through various measuring tools and come to the conclusion that it is in particular the "monsterjoin" in the first sentences of my mail that takes time. The odd part is that it doesn't seem like the machine sets the limit. Does anyone know what could make a difference?
I thank you very much for taking the time and effort and looking and hopefully answering my mail
Regards
Daniel
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
