Managed disaster recovery

Managed disaster recovery (DR) replicates your Azure Databricks deployment to a secondary region so you can recover from a regional outage in minutes. Azure Databricks manages the replication pipeline, the state of the replicated catalogs in the secondary, and the failover process. You do not write or maintain replication scripts.

For the manual approach to disaster recovery, including general DR concepts and best practices, see Disaster recovery.

Important

Managed DR is gated. Apply for access through your Azure Databricks account team. Azure Databricks enables managed DR on your account after you are accepted.

What is managed disaster recovery?

Managed DR uses the workspaces and metastores you already operate. You bring two Azure Databricks workspaces, one in your primary region and one in your secondary region, and a metastore in each region. You group them in a failover group, and managed DR then:

  • Replicates the categories you choose from the primary to the secondary on a continuous schedule. You can opt into either or both categories: Unity Catalog metadata and managed table data, and workspace assets such as notebooks, jobs, SQL warehouses, clusters, and access control lists (ACLs).
  • Provides an optional stable URL, a single connection string that points to the current primary, so clients keep working after failover without reconfiguration. Infrastructure-as-code tools use the matching stable workspace ID.
  • Lets you trigger failover at any time, for a DR test or a real outage.

Workspace asset IDs are preserved across regions, so URLs that reference a workspace asset by ID still resolve after failover.

Managed DR does not create or replicate your cloud storage. You create the storage in the secondary region, register it in Unity Catalog, and allow network access between the two regions before you create a failover group. See Set up managed disaster recovery.

Requirements

For the requirements to set up managed DR, see Requirements.

What gets replicated

Managed DR can replicate the following on every replication cycle.

Unity Catalog metadata and data

Object Notes
Unity Catalog managed tables in Delta Lake Data and metadata. See Limitations.
External tables Metadata only. See Limitations.
External volumes Metadata only.
Views Views that reference other catalogs require the view owner to have access to those objects in the secondary. See Reference.
Functions Objects that a function references must also be replicated. See Reference.
Permission grants Ownership of replicated objects is transferred to match the primary. See Ownership of replicated objects.
Streaming tables In Private Preview. Contact your Azure Databricks account team for more information.
Materialized views In Private Preview. Contact your Azure Databricks account team for more information.
Row filters, column masks, and attribute-based access control (ABAC) policies In Private Preview. Contact your Azure Databricks account team for more information.
Foreign catalogs In Private Preview. Contact your Azure Databricks account team for more information.

The catalog's isolation mode is also replicated. If the source catalog is open, its replica is open. If the source catalog is isolated and bound to the primary workspace, the replica is isolated and bound to the secondary workspace.

Workspace assets

Workspace assets are replicated along with their ACLs.

Object Notes
Notebooks Notebook source only.
Jobs Job schedules in the secondary are paused.
SQL warehouses Replicated in STOPPED state.
Clusters Replicated in TERMINATED state. Also includes compute policies, instance profiles, and global init scripts.
Instance pools Capacity reservations are region-specific and must be configured in the secondary.
If your failover group was created before September 17, 2026, contact your Azure Databricks account team to add instance pools to your failover group.
Draft AI/BI dashboards After failover, publish draft dashboards.
Files Workspace files only.
Folders Does not include Git folders.
Queries Includes queries created and saved in the SQL editor.
Libraries Includes cluster-installed libraries.
Notification destinations Destinations used for job and alert notifications.
Genie Agents After failover, refresh prompt matching in each Genie Agent to improve answer accuracy.
If your failover group was created before October 30, 2026, contact your Azure Databricks account team to add Genie Agents to your failover group.
Lakeflow pipelines In Private Preview. Contact your Azure Databricks account team for more information.

Resources created by managed DR

When you create a failover group, managed DR provisions auxiliary Unity Catalog resources that the replication pipeline uses to copy data between regions. In both the primary and secondary metastores, managed DR creates:

  • A connection that points to the workspace in the other region.
  • A foreign catalog for each replicated catalog. The foreign catalog references the corresponding catalog in the other region.

These resources appear alongside your own catalogs in Catalog Explorer. You can identify them by their comment, which notes that Azure Databricks disaster recovery created and manages them.

Important

By default, only a metastore admin can modify or delete these resources. While a failover group is active, do not delete the connections or foreign catalogs that managed DR creates. Deleting either one breaks replication for the failover group.

Ownership of replicated objects

When managed DR creates a replicated securable (catalog, schema, table, view, function, or volume) in the secondary, the initial owner is the Azure Databricks service principal that runs the replication, because Unity Catalog assigns ownership to the identity that creates the object. Managed DR then transfers ownership of the replica to match the owner of the corresponding securable in the primary.

If the owner of a securable in the primary is a user who has been deleted from the account, managed DR can't replicate the owner. For most securables, replication continues and the replica keeps its current owner: the Azure Databricks service principal if managed DR creates the securable after the user was deleted, otherwise the deleted user. Functions are the exception. If managed DR has to create or update a function, replication of that function fails with DR_SOURCE_ENTITY_ERROR.OWNER_UNRESOLVABLE, and the failover group's recovery point objective (RPO) doesn't advance until you fix it. A function that was already replicated and hasn't changed keeps replicating. To resolve this, assign a valid owner to the securable in the primary and let managed DR replicate it.

Note

If you plan to reuse a secondary workspace in a new failover group, verify that it meets the configuration prerequisites and delete the following before you create the new failover group:

  • Catalogs with the same names as the catalogs being replicated.
  • Foreign catalogs with names that start with r_. See Resources created by managed DR.
  • Leftover external tables and external volumes whose storage paths overlap the storage paths of the tables and volumes being replicated.

Limitations

Managed DR has the following limitations:

  • Not replicated: managed volumes, Unity Catalog and workspace secrets, ML models, Model Serving endpoints, AI Search indexes, OpenSharing shares, Genie Agent conversation history, published AI/BI dashboards (drafts replicate), and Structured Streaming outside Lakeflow pipelines. Tables with Delta Lake archival support (delta.timeUntilArchived) are flagged as Failed to replicate in the system table, and these failures hold up the RPO until you remove the resource from the failover group's scope. Metadata replication is not supported for non-Delta Lake external tables that use partition metadata.
  • Tables with row filters or column masks and ABAC-tagged resources are flagged as Failed to replicate in the system table, and these failures hold up the RPO until you remove the resource from the failover group's scope. Replication for these resources is in Private Preview. Contact your Azure Databricks account team for more information.
  • Row filters, column masks, and ABAC policies that use a function from a different catalog cannot be replicated, even when both catalogs are in the failover group. These resources are flagged as Failed to replicate in the system table with DR_UNSUPPORTED_FEATURE.CROSS_CATALOG_POLICY_FUNCTION, and the failure holds up RPO until the function is in the same catalog as the table or policy that uses it.
  • Managed table writes from external engines have limited detectability. Managed DR detects changes to Unity Catalog managed tables from writes made by Azure Databricks compute. Writes to a replicated managed table from an external (non-Azure Databricks) engine through open APIs, such as the Iceberg REST Catalog, might not be detected, so those writes might not replicate to the secondary and can be lost during failover. For tables you replicate with managed DR, write through Azure Databricks compute.
  • Secondary in-scope catalogs are read-only. Read-only applies only to replicated entities. You can still set up your own replication for securables outside the managed DR scope. However, you cannot run compute on the secondary workspace while managed DR is enabled, which limits operating a do-it-yourself replication pipeline there.
  • Renaming a Unity Catalog securable triggers a delete and recreate in the secondary. For managed tables, the rename re-replicates the table data on the next cycle. Avoid renaming during steady-state replication.
  • Shallow clones of tables in a failover group's scope can become unreadable after a failover. To keep a copy usable across failovers, use a deep clone. Otherwise, recreate the affected shallow clones after the failover.
  • UNDROP is not propagated to the secondary.
  • For tables without in-commit timestamps, time travel by timestamp can return the wrong version. On these tables, Delta Lake takes each commit's timestamp from the commit file's modification time, and replicated commit files carry the time they were copied rather than the original commit time. As a result, time travel queries and change data feed reads by timestamp can resolve to a different version than on the source table, without raising an error. This affects the secondary as soon as a table replicates, and the primary after a failover or failback. Time travel by version is not affected. To avoid this, enable in-commit timestamps on these tables before you set up replication. See Prepare the workspaces.

Additional resources

Resource Description
Set up managed disaster recovery Prepare the secondary region, configure storage and network access, and create a failover group.
Manage, fail over, and troubleshoot managed disaster recovery Monitor replication, fail over and fail back, and resolve replication errors.