Cloud Spanner Cloud Instance

The Cloud Spanner instance has the following properties:

To change the migration team size, open the Migration Team Size dropdown menu at top of the page, and select value from the common team sizes in the list.

Add Instance

To add Cloud Spanner target instance:

1. In the Cloud Spanner tab, click Create.

2. In the left part of the instance configuration window, select the checkboxes of the sources to migrate to the target instance.

3. In the Cloud Instance Information panel in the instance configuration window:

   1. Select the Instance Group from the dropdown

   2. Enter an Instance Name for the target instance

   3. Select the GCP Spanner option in the Target Platform dropdown  

4. In the Cloud Instance Information panel in the instance configuration window, configure the instance:

   • Select Regional or Multi-Region configuration

   • Select the Compute Units type

   • Set the Processing Units or the Nodes value

   • Assign the Region

   • Select the Committed Usage period

   • The Target Storage Utilization value is automatically calculated based on the cumulative size of the sources selected, with a percentage padding to allow for growth and operation.

   • Set the Backup size 

5. Click the Confirm button to save the configuration or the Cancel button to discard it.

Compute Unit and Sizing

Cloud Spanner has characteristics and limits that affect the Processing Units or Nodes count settings.

The initial value is automatically calculated and assigned in an attempt to adhere to the following limits:

• Maximum database storage allowed per node of 4 TiB.

• Maximum number of databases allowed per node of 100.

• Maximum read operations per second allowed per node of 10K ops/sec (14K if multi-region).

• Maximum write operations per second allowed per node of 2K ops /sec (1.8K if multi-region).

• Region-specific maximum node count per project.

To accommodate real-world operational scenarios, additional resources are allocated to the underlying required resources to allow natural growth and spike usage.

mMC collects the data sizes and various object counts during the source scanning.

Workload data is collected from AWR reports in Oracle, and via Extended Scan mode in mMC for other database engine types.

Attempting to under-provision the target instance concerning the sources selected to migrate into it produces an error. 

For external documentation related to sizing and limitations, see:

• Compute capacity, nodes, and processing units

• Quotas & limits

• Regional and multi-region configuration