Cassandra supports incremental backups, and the recommendation is to do a full backup, and then enable incremental backups. From then on, backups will be incremental.
If we're starting off with a new cluster, with no data, is it possible to either:
a) Have incremental enabled from the get go. b) Run a full backup as part of the install script on each node, and then enable incremental backups on each. In other words: on node 1, install cassandra, run nodetool snapshot, enable incremental, then on node 2, install cassandra, run nodetool snapshot...and so on.
Thinking about the best way to get incremental going with puppet.
