El día 22 de diciembre de 2009 13:44, Miguel Angel Nieto <cor...@miguelangelnieto.net> escribió: >> It depends a lot on how you plan to coordinate the db servers >> (sharding, replication, ndb), the kind of applications you are going >> to deploy and how much scability you need. > > Thank you. I have read about LVS and keepalived but I can't see the > difference between them. Are they the same thing? I want the load > balancing for my replicated servers. I suppose that LVS can't > distinguish between inserts and selects (to send queries to the master > o slave server).
I do not know about keepalived, but it seems to provide the same service that heartbeat does: user-level awareness of the failure of a machine or a service. LVM, however, does kernel module, ip-level, load balancing and automatic failure clustering. The big confusion with those apps is that most of them are extensible and combinable to achieve the same goal. About the read/write balancing (which is a good question, and that is why I asked what you were intending to do on the other end), I have seen it done **only at application level**, not transparently, because of the problems derived from asynchronous replication (lag between master writes and slaves see the data). I am sorry I cannot help you, but please, share here if you found something useful AND with good performance. Merry Xmas! -- Jaime Crespo MySQL & Java Instructor Warp Networks <http://warp.es> -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/mysql?unsub=arch...@jab.org