tag:blogger.com,1999:blog-14455177.post112927545016167635..comments2024-03-07T18:57:25.977+01:00Comments on Mikael Ronstrom: Default Partitioning of MySQL Cluster 5.1Mikael Ronstromhttp://www.blogger.com/profile/07134215866292829917noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-14455177.post-1168917093300810402007-01-16T04:11:00.000+01:002007-01-16T04:11:00.000+01:00hello, Mikael Ronstrom: I am a graduated student o...hello, Mikael Ronstrom:<BR/> I am a graduated student of china.<BR/>When i debug mysql source code ,I have many questions about mysql ndb cluster partition and query paralle. I want to communicate with you quickly.Can you give me your email.<BR/>my email:fzb0717@sohu.comAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-14455177.post-1132858123898811602005-11-24T19:48:00.000+01:002005-11-24T19:48:00.000+01:00Mikael, this information is great!-johanMikael, this information is great!<BR/><BR/>-johanJohan Anderssonhttps://www.blogger.com/profile/09202425156709885026noreply@blogger.comtag:blogger.com,1999:blog-14455177.post-1129345457463740092005-10-15T05:04:00.000+02:002005-10-15T05:04:00.000+02:00I read through that, and am not quite sure if it s...I read through that, and am not quite sure if it solves the one thing I'd really like from partitioning. I want to be able to specify which column to balance the data by. The reason for this, is I want all data related to one user on one server. Take my Messaging system as an example. I run alot of queries like "SELECT * FROM msgs WHERE userid = ?", and have an index on userid. Why not guarantee that all of that data is on one server, to reduce the load on the other servers and to reduce the number of connections. At the moment, since it isn't on the primary key (an auto_increment which is unrelated to the userid), that query would use all servers. The ability to add servers/partitions without taking the database down would also be great for scalability.Anonymousnoreply@blogger.com