Migrate to managed disaster recovery

This page describes how to migrate from BigQuery cross-region replication to BigQuery managed disaster recovery.

Overview

BigQuery cross-region replication (CRR) and managed disaster recovery (DR) are both features designed to enhance data availability and disaster recovery capabilities. However, they handle regional outages in different ways. CRR doesn't allow promoting the secondary replica if the primary region is unavailable. In contrast, DR offers more comprehensive protection by allowing a failover to the secondary replica even if the primary region is unavailable. With CRR, only storage is replicated, while with DR both storage and compute capacity are replicated.

The following table describes the feature capabilities of CRR and DR:

Migration implications

The following sections provide an overview of the cost and capability changes that occur when you migrate to DR.

Cost implications

Consider the following cost implications when you migrate from CRR to DR:

• DR only supports write access from the Enterprise Plus edition, which incurs higher compute costs. You can read from any capacity model, so read costs for existing jobs don't change.

• DR uses Turbo replication, which incurs additional costs depending on the region pair.

• Storage prices are the same for both CRR and DR.

For more information about pricing, see Pricing.

Capability implications

Consider the following capability implications when you migrate from CRR to DR:

• DR only supports failover at the reservation level. Any existing jobs that rely on dataset-level failover fail.

• Only Enterprise Plus edition queries can write to the dataset once it is attached to the DR reservation. Any existing write jobs that don't use an Enterprise Plus edition for their compute capacity fail.

Before you begin

Before you begin the migration, familiarize yourself with the concepts in cross-region replication and managed disaster recovery.

To migrate to DR, you must have the following prerequisites:

• You have an active Google Cloud project with BigQuery enabled.

• You have created and replicated datasets with CRR.

• The datasets have the same primary and secondary locations you want to use for DR.

• You have the necessary permissions to work with DR. For more information about permissions, see Before you begin.

Migrate from CRR to DR

The following sections describe how to migrate your datasets from CRR to DR. It assumes you have already configured your datasets for CRR.

Create a failover reservation

To enable disaster recovery, you must create a failover reservation in the primary region. Configure your reservation with the appropriate primary and secondary region. The primary and secondary regions should match the regions of all the CRR datasets you intend to migrate to DR. To create a failover reservation, choose one of the following options:

Attach the dataset to the reservation

Once you've created the failover reservation, attach your cross-region dataset, or datasets, to the reservation. This enables failover for all of the attached datasets. To attach the dataset to the reservation, choose one of the following options:

Verify the configuration

To verify the status of your configuration, query the INFORMATION_SCHEMA.SCHEMATA_REPLICAS view.

PROJECT_ID.`region-REGION`.INFORMATION_SCHEMA.SCHEMATA_REPLICAS[_BY_PROJECT]

Verify that the datasets are attached to the correct reservation in the correct regions.

• Optional: PROJECT_ID: the ID of your Google Cloud project. If not specified, the default project is used.
• REGION: any dataset region name. For example, `region-us`.

Examples

The following example walks you through the steps to migrate from CRR to DR with practical examples using GoogleSQL. For this example, assume the following:

• You are working in a project named myproject.

• You have already created a dataset named mydataset and configured it with CRR.

• The primary region of mydataset is us-central1 and the secondary region is us-west1.

To begin migrating your dataset to DR, first create a reservation with the Enterprise Plus edition. In this example, the name of the reservation is myreservation.

CREATE RESERVATION `myproject.region-us-central1.myreservation`
OPTIONS (
  slot_capacity = 0,
  edition = ENTERPRISE_PLUS,
  autoscale_max_slots = 50,
  secondary_location = 'us-west-1');

Once the reservation is created, you can then attach the dataset to the reservation. The following example attaches the dataset to the reservation:

ALTER SCHEMA
  `myproject.mydataset`
SET OPTIONS (
  failover_reservation = 'myproject.myreservation');

Then, verify that the dataset has been successfully attached.

SELECT
  failover_reservation_project_id,failover_reservation_name,
FROM
 `myproject`.`region-us-west1`.INFORMATION_SCHEMA.SCHEMATA_REPLICAS
WHERE
 schema_name='mydataset';

The results of this query should look similar to the following:

+---------------------------------+---------------------------+
| failover_reservation_project_id | failover_reservation_name |
+---------------------------------+---------------------------+
| myproject                       | myreservation             |
| myproject                       | myreservation             |
+---------------------------------+---------------------------+

What's next?

• For more information about cross-region replication, see Cross-region dataset replication.

• For more information about managed disaster recovery, see Managed disaster recovery.