You should look into the HA (High availability) for mysql. It uses mounted partition space for cluster and can be used with mysql, innodb tables.
ndbcluster tables can work against you if your using lots of table joins or yoru database gets very large.
The way clustering works on joins in the API has to re-assemble the table on the fly and then eliminate the unwanted database before it can perform the query. This can severely slow down the query process.
Only consider mysql clustering if your going to be doing simple queries on an average database. MySQL even though it is an in RAM product. Can't keep up with queries when complex joins are being used.
RedHAT enterprise 5 has just been released which is something that is supposed to solve these issues by clustering together machines and using resources wisely. For both HA and failover. Have to tried working with it yet myself but the concept sounds interesting.
Quote:
Originally Posted by stefano65
Hello everybody,
I would like to clustering mysql using red hat cluster suite with a shared storage ( I do not want to use mysql native clustering because it is a all-in-RAM database and my database will grow very much) .
Is there a tutorial or a whitepaper that can show me how to do, explaining me, for example, if I can use innodb engine for mysql table or I have to use ndb engine? Does Clustering mysql with red hat cluster suite allow only a failover configuration or also a load balancing configuration is permitted?
Thanks a lot for any help.
Stefano
|