Migrating or renaming a cluster
The information on this page is intended for the following types of scenarios:
-
Migrating a cluster, including transitioning an EC2 cluster to Amazon virtual private cloud (VPC), moving a cluster, or upgrading from an early version cluster to a recent major version.
-
Renaming a cluster. You cannot change the name of an existing cluster; you must create a new cluster and migrate your data to the new cluster.
The following method migrates a cluster without service interruption and ensures that if a problem occurs in the new cluster, you still have an existing cluster as a fallback.
Procedure
-
Set up and configure the new cluster as described in Initializing a DataStax Enterprise cluster.
If you’re not using vnodes, be sure to configure the token ranges in the new nodes to match the ranges in the old cluster. See Initializing single-token architecture datacenters.
-
Set up the schema for the new cluster using CQL.
-
Configure your client to write to both clusters.
Depending on how the writes are implemented, code changes may be required. Be sure to use identical consistency levels.
-
Ensure that the data is flowing to the new nodes so you won’t have any gaps when you copy the snapshots to the new cluster in 6.
-
Snapshot the old cluster.
-
Copy the data files from your keyspaces to the nodes.
-
You can copy the data files to their matching nodes in the new cluster, which is simpler and more efficient, if:
-
You are not using vnodes.
-
Both clusters use the same version of DataStax Enterprise (DSE).
-
The node ratio is 1:1.
-
-
If the clusters are different sizes or if you are using vnodes, use the sstableloader (sstableloader).
-
-
You can either switch to the new cluster all at once or perform an incremental migration.
For example, to perform an incremental migration, you can set your client to designate a percentage of the reads that go to the new cluster. This allows you to test the new cluster before decommissioning the old cluster.
-
Ensure that the new cluster is operating properly and then decommission the old cluster. See Decommissioning a datacenter.
