How to do a Live Migration of a vanilla Cassandra cluster to a Datastax Enterprise Cluster

Question

We currently have a Cassandra cluster up and running in a production environment using Cassandra 3.3.0. We want to switch this to a DSE cluster without having any down time.

We have a tool for migrating the schema already, but what is the best way to migrate the data from our Cassandra cluster to the new Datastax cluster without any sort of down time?

If you were switching clusters with regular Cassandra, you could merely join Nodes from the new cluster to the currently running cluster, then shutdown the old nodes after all of the new nodes have been added. This is nice because it requires no additional code to migrate the data, is there any solution similar to this available for DSE?

Answer 1

Normally you would install DSE to match your version of Cassandra around the cluster. However I noticed you're listing Cassandra 3.3.0, is that correct? I checked the Cassandra release notes but couldn't find that version.

Assuming you mean 3.0 then you can install DSE 5.0 on your nodes without any need for schema migration since it'll be the same major release of Cassandra.

Install docs should all be covered here:

https://docs.datastax.com/en/latest-dse/datastax_enterprise/install/installTOC.html

If however you don't want to go that route you could also standup a new DSE cluster and stream your data across using the sstableloader tool:

https://docs.datastax.com/en/cassandra/3.0/cassandra/tools/toolsBulkloader.html

Of course it goes without saying you should always test first before doing anything in production!