3

How does the newly released Percona XtraDB Cluster compare with MySQL Cluster?

http://www.mysqlperformanceblog.com/2012/01/09/announcement-of-percona-xtradb-cluster-alpha-release/

archmeta
  • 1,107
  • 4
  • 17
  • 29

1 Answers1

3

Simply the Percona XtraDB Cluster is MySQL 5.5 with the Galera replication (synch multimaster) + other patches. That means that you still keep your data in disk + ram cache where MySQL Cluster keeps all data in RAM and saves to disk. In short, if your dataset is bigger than your RAM you need Percona's cluster.

aemadrid
  • 229
  • 1
  • 2
  • 7
  • So to confirm, xtradb cluster will NOT scale writes, just like any other multi-master sync replication mechanism, and one must use mysql ndb cluster to scale writes? – archmeta Jan 15 '12 at 22:06
  • Not sure what you mean by scaling writes but I do see XtraDB Cluster more like a really nice HA solution for me where I don't need that much RAM like with MySQL cluster. – aemadrid Jan 19 '12 at 20:22
  • 1
    Note that MySQL Cluster does give you the option of storing your all or a sub-set of your non-indexed columns on disk rather than in memory and so you can scale beyond your available RAM. It's also worth noting that MySQL Cluster allows you to scale out linearly - both reads and writes - and provides very high levels of availability/fault tolerance. – Andrew Morgan Feb 29 '12 at 15:12