2 Prepare to Migrate Oracle Data Hub Cloud Service to Oracle Cloud Infrastructure

You can perform a live migration of your existing Oracle Data Hub Cloud Service cluster to DataStax Distribution of Apache Cassandra or DataStax Enterprise on Oracle Cloud Infrastructure IaaS by deploying new nodes and syncing the data. At a high level, this is accomplished by upgrading the existing nodes in Data Hub, joining new nodes in Oracle Cloud Infrastructure to the Data Hub cluster as a new data center, and by running a rebuild command or loading snapshot files.

Topics

• Options to Deploy DDAC/DSE on Oracle Cloud Infrastructure
• Size your Oracle Cloud Infrastructure Instances
• Check Data Hub Node Configuration
• Upgrade your Data Hub Nodes
• Prepare your Network

Options to Deploy DDAC/DSE on Oracle Cloud Infrastructure

• The Oracle Cloud Infrastructure team along with the DataStax Solutions team has built Terraform deployment templates to automate deployment of DDAC and DSE on Oracle Cloud Infrastructure. These templates deploy all required infrastructure resources (virtual cloud network, subnet, security list, compute instances, storage) as well as DDAC/DSE. Please reach out to the Oracle Sales team to gain access to the Terraform template and its documentation.
• Both DDAC and DSE are available for download from DataStax's website. The customer is required to obtain a license from DataStax when installing the DDAC or DSE. Follow DataStax's documentation for installing and running of both solutions.

Size your Oracle Cloud Infrastructure Instances

The number and size (in terms of cores, RAM, and block storage) of your new nodes in Oracle Cloud Infrastructure should match those in Data Hub, choosing core number first.

• For example, if you deployed your Data Hub cluster using an OC6 (8 OCPU/60 GB RAM) shape you would deploy the same number of VM.Standard2.8 (8 OCPU/120 GB RAM) instances. A full list of Oracle Cloud Infrastructure shapes can be found here.
• The block storage volume for each node should be sized at 200% of the data contained on each Data Hub node to allow for growth and compaction operations. Disk sizing is discussed in detail here.

Check Data Hub Node Configuration

The cassandra.yaml file (default location: /etc/cassandra/) on all nodes should set the endpoint_snitch property to GossipingPropertyFileSnitch.

Upgrade your Data Hub Nodes

For both migrations, all existing Data Hub nodes need to be upgraded to either DDAC or DSE. This is a mandatory step because some operations required cannot be performed on a cluster with mixed versions. Instructions and a compatibility matrix are available here.

Prepare your Network

The new cluster nodes will be deployed into an Oracle Cloud Infrastructure VCN. These new nodes should either be deployed in a:

• Private subnet of an Oracle Cloud Infrastructure VCN that has been connected with your Oracle Cloud Infrastructure Classic network by following these instructions. Also in the VCN's security list port 7000 (port 7001 if using inter-node SSL) need to be open between the Data Hub andOracle Cloud Infrastructure CIDR blocks as described here.
• Public subnet. On each Oracle Cloud Infrastructure node set the broadcast_address property in cassandra.yaml to the node's public IP. The provided terraform templates will configure this property if the nodes are deployed to a public subnet. It should be noted that after the migration you can manually set this to the node's private IP. Also in the VCN's security list port 7000 (port 7001 if using inter-node SSL) need to be open between the Data Hub and Oracle Cloud Infrastructure CIDR blocks as described here.