Notes from the MySQL Cluster High Availability Features talk at the 2004 MySQL User's Conference...
Redundancy between nodes via heartbeat. Redundancy between clusters via replication. In other words, NDB provides local and global redundancy. System recovery for a full-scale shutdown. Hot backup and restore. The architecture in general was designed to eliminate single points of failure.
Lots of diagrams illustrating recovery in various failure scenarios that I can't ASCII copy too easily. Doh!
Posted by jzawodn at April 14, 2004 02:51 PM
Hmm, did they happen to have any live demonstrations of this going on or was it all just diagram after diagram? :-) This kind of high availability rocks, just from the sound of it..
It'd be way cool to spend a week learning about stuff like this..
Notice that, however, the new MySQL cluster has some fishy points.
It doesn't support BLOB columns or FULLTEXT indexes, for instance.
It requires at least an amount of RAM twice the size of your database + 10%.
It is still in alpha stage (it depends on the 4.1 version).
What type of hardware servers are ideal for MySQL clusters (eg, quad Xeon, blade, or ??)?
it's too bad you didn't have a cameraphone or a camera with you and you could take snaps of the slides!
Hello,
There are some scripts that are available to do monitoring of a failover in setups like the one listed here :
http://www.karkomaonline.com/article.php?story=2004012416185184
The scripts to monitor the mysql service and fail over the cluster if it fails to get a mysqladmin ping response. The script to do this can be found here :
Nice information for developers. Clusters are very helpfull.