Forum Discussion

fabricpribeiro's avatar
fabricpribeiro
Icon for Post Patron rankPost Patron
6 months ago
Solved

Disaster Recovery in Fabric

Dears,   I have the below architecture in fabric  (figure 1)   I need to create the strategy for DR which I will than need to defend at TDA   Not sure very well which topics to cover. I made a ...
workspaces
  • v-echaithra's avatar
    v-echaithra
    6 months ago

    Hi fabricpribeiro ,

    Fabric DR is different from traditional backup/restore.

    > Fabric protects data automatically through OneLake geo-replication (capacity-level, Microsoft-managed, no performance impact).
    > This only safeguards the data, it does NOT replicate workspaces, pipelines, or semantic models.

    So DR in Fabric = Redeploy + Rehydrate, not restore.

    In a regional outage:
    Microsoft ensures the data is available in the paired region.
    Customer redeploys the Fabric environment from Git/source control to resume operations.

    Minimum recommended setup

    Enable Geo-redundancy on Production capacity
    Use Git Integration for all Fabric artifacts
    Keep a redeployment runbook/scripts ready
    Ensure source data can be reprocessed if needed
    Test DR periodically.

    Hope this helps.
    Thank you

fabricpribeiro's avatar
fabricpribeiro
Icon for Post Patron rankPost Patron
6 months ago

Sorry, just a couple more questions , please:

 

1 - If I choose geo-redundacy it will be at capacity level that this option is available, right?

2 -  The Geo.redundancy means that the data will be replicated inside only Europe regions? or can it be replicated to outside of Europe if my capacity wass created at Western Europe?

3 - If I decide not to activate this option, what do we have as default in fabric for DR? do we have protection at least at data center level with the same data being replicated inside the same data center to multiple different redundant discs?

 

Thanks,

 

Pedro

 

  • v-echaithra's avatar
    v-echaithra
    Icon for Community Support rankCommunity Support
    6 months ago

    Hi fabricpribeiro ,

    Thank you for your response.

    Yes, Geo-redundancy (OneLake disaster recovery) is enabled at the Fabric capacity level, not at workspace or item level.
    Reference: OneLake Disaster Recovery and Data Protection - Microsoft Fabric | Microsoft Learn

    Replication follows Azure regional pairing rules. If your capacity is in West Europe, the paired region is North Europe. Azure regional pairing always stays within the same geography (Europe > Europe) for compliance and data residency reasons.
    It will not replicate outside Europe. The paired region is automatically selected by Azure and cannot be changed.

    Even without cross-region Geo-redundancy, Fabric still provides Intra region durability, Redundant storage within the same region, Data replication across fault domains inside Azure datacenters, Encryption at rest and in transit. This protects against Disk failures, Rack failures, Datacenter-level hardware issues inside the same region
    However, it does not protect against a full regional outage (e.g., West Europe unavailable).
    Even without enabling Geo-redundancy, Microsoft Fabric still provides built-in resiliency within the same region. However, this is High Availability, not full Disaster Recovery.

    Hope this helps.
    Thank you.