Microsoft Fabric BCDR storage

Question

As using Azure price calculator for Fabric SKU, I wonder if I select hot, cool and cold storage as BCDR OneLake hot, cool and cold storage then the whole of those hot, cool and storage will be replicated from this primary region to another region? And for the secondary Fabric SKU in DR region, I don't need to set any storage, because the storage from primary region will be available in OneLake to be accessed for the secondary SKU?

Answer 1

Selecting hot, cool, and cold storage for OneLake in the Azure Pricing Calculator only models cost for those tiers in the primary region. It does not represent three separate, independently configured BCDR storages.

For Microsoft Fabric and OneLake:

1. What gets replicated with DR enabled
   • When the disaster recovery (DR) capacity setting is enabled for a Fabric capacity, OneLake data in that capacity (Lakehouse and Warehouse data) is geo‑replicated from the primary region to its Azure‑paired secondary region.
   • This replication applies to the OneLake data itself, regardless of whether it is stored in the hot, cool, or cold storage tier. All three tiers are online and supported; DR operates at the OneLake data level, not per‑tier.
   • DR is configured at the capacity level, not per workspace or per storage tier.
2. Do hot/cool/cold tiers get duplicated in the DR region?
   • Yes in effect: enabling DR means the OneLake data (including data in hot, cool, and cold tiers) is stored in two geographic regions so that it is geo‑redundant.
   • The cool and cold tiers still retain their characteristics (lower storage cost, higher access cost, minimum retention periods) and, when accessed, will incur the documented BCDR read/write and retrieval CU consumption for those tiers in DR scenarios.
3. Do you need to configure storage again on the secondary Fabric capacity?
   • The DR toggle only affects OneLake data for the capacity where it is enabled. It does not require a separate storage configuration on a secondary capacity for that same data.
   • After a regional failover, OneLake data is accessed via the global endpoint in the secondary region; the replicated OneLake data is available there without separately provisioning another OneLake store for that same capacity’s DR copy.
   • If a separate Fabric capacity is deployed in the DR region for other workloads (for example, independent warehouses or pipelines), that capacity’s own OneLake usage and storage tiers are configured and billed independently. DR for one capacity does not automatically cover data stored outside OneLake or in other capacities.
4. Cost and CU considerations for DR and tiers
   • When DR is enabled, write operations consume higher capacity units (CUs) for BCDR operations (for example, OneLake BCDR Write). Rates differ by hot/cool/cold tier.
   • For cool and cold tiers, data retrieval from OneLake incurs additional CU consumption per GB read, with higher rates for colder tiers.
   • These costs apply to the replicated data as it is written and read under DR.

Summary: enabling DR for a Fabric capacity causes OneLake data (across hot, cool, and cold tiers) in that capacity to be geo‑replicated to the paired region. The DR copy is part of the same OneLake DR setup; it does not require separate storage configuration for that same data on a secondary capacity. Any additional capacities in the DR region are independent and must be configured separately if they host their own data.

---

References:

• OneLake storage tiers (preview)
• OneLake compute and storage consumption
• Disaster recovery and data protection for OneLake
• Reliability in Microsoft Fabric
• Disaster recovery scenario details for an Azure data platform