Howto: Build and scale a Cassandra cluster in five minutes

Part of a series:

memcached - http://blog.gridcentriclabs.com/2010/07/how-to-build-10-core...

Hadoop - http://blog.gridcentriclabs.com/2010/07/howto-build-hadoop-c...

An early take on Hudson / Continuous Integration - http://blog.gridcentriclabs.com/2010/07/elastic-build-system...

As always, we'd love any feedback the HN community might have!

This is MILES better than when I attempted to create a virtualized Oracle RAC 10g cluster. Never before have I cursed a company's name so much.

This looks downright pleasant to configure.

Thanks! The low-effort VM clone (essentially fork() for Xen VMs, and only takes 1-2 seconds even for dozens of clones) makes deploying a bunch of almost-identical machines extremely easy.

We've also toyed around with other cluster-scale platforms (see my earlier post) and have a MySQL Cluster HOWTO in the works, but the clustered SQL databases are decidedly more heavyweight - MySQL Cluster doesn't even allow removal of cluster nodes once they're added (without dropping all the clustered tables). Not sure if Oracle RAC has that capability - if so, we might be able to put together a fun demo.

hfdghf hgfghfgh