EDB Brings Postgres-Proven Disaster Recovery to Petabyte-Scale WarehousePG Where Downtime Is Not Possible

For banks and regulated enterprises running analytics around the clock, a warehouse recovery measured in days is a business risk. EDB Postgres^® AI for WarehousePG brings it down to minutes and lets enterprises keep control over where that data runs.

WILMINGTON, Del., Aug. 20, 2026 /PRNewswire/ -- Today, EnterpriseDB (EDB), the leading sovereign data and AI company, announced the general availability of Disaster Recovery for WarehousePG, adding mission-critical capabilities to the analytical foundation of EDB Postgres AI (EDB PG AI). Customers now have a supported path to protect petabyte-scale data and restore warehouse operations across infrastructure boundaries. Enterprises can keep a recovery environment current wherever they choose—on-premises, in the cloud, or in a hybrid deployment—and decide when to bring it online.

A warehouse outage is no longer a reporting problem. It is an operational event. AI agents, analysts, and business teams increasingly depend on live analytical data to make decisions and keep critical services moving. Yet at petabyte scale, a nightly backup can leave hours of data exposed, while a logical restore can take days and compete with production workloads along the way. For enterprises in regulated and mission-critical environments, that gap puts recovery objectives, customer commitments, and data sovereignty at risk.

WarehousePG now extends EDB PG AI's converged analytics capabilities with a critical layer of resilience for enterprise-scale analytical workloads. Organizations can keep a separate analytical environment ready for recovery while retaining control over where the standby runs, how frequently restore points are created, and when the standby is promoted.

"At petabyte scale, a recovery process that takes days is not a recovery plan, it is unhedged business risk. WarehousePG brings the full distributed cluster back to a consistent point in time without forcing customers to trade production performance for protection. That moves disaster recovery from a best-effort exercise to a capability teams can test, measure, and trust," said Max Romanenko, chief engineering officer, EDB.

From two days to 16 minutes

In one banking deployment supporting hundreds of terabytes of business-critical data and a 24x7 analytical pipeline, the existing logical backup and recovery process had outgrown the warehouse. Full backups took nearly two days, restores took just as long, and backup jobs ran through most of the day. With Disaster Recovery for WarehousePG, the customer established hourly restore points on cost-optimized recovery infrastructure, without changing the application architecture.

• Approximately 180x faster recovery: Restore-point replay averaged 16 minutes, compared with a process that previously required nearly two days.
• Four-times faster full backups: Full backup time fell from 44 hours to approximately 11 hours.
• Nearly 99% less query lock time: Physical backup sharply reduced the lock contention imposed by the previous logical backup process.
• No application redesign: The customer achieved the new recovery model without changes to the application architecture above the warehouse.

Results reflect a specific customer deployment; actual outcomes vary by environment, workload, and configuration.

Postgres-proven recovery, engineered for MPP scale

Because WarehousePG runs on Postgres, disaster recovery is familiar to Postgres teams — now applied across a full standby cluster for warehouse scale. The coordinator and every primary segment recover together to a single point in time, so the standby cluster comes up as one consistent warehouse. EDB has built and supported mission-critical Postgres disaster recovery for years. Disaster Recovery for WarehousePG brings that expertise to petabyte-scale analytics, using open source Barman and the Postgres write-ahead log (WAL) to coordinate recovery across the distributed cluster, giving enterprises a faster and supported path back to operations without adding a separate proprietary stack.

WarehousePG's built-in high availability uses segment mirrors and a standby coordinator to protect against component failures within the cluster. Disaster Recovery for WarehousePG addresses the next level of risk—the loss of an entire site—by enabling the full cluster to be recovered in a separate environment.

WarehousePG Enterprise Manager already gives operations teams a unified place to manage cluster health, query diagnostics, and resource utilization. With the impending September version update, Disaster Recovery configuration will be included as well, unlocking new value for customers:

• Recovery aligned to business objectives: Teams can set restore-point frequency, continuously archive WAL, and validate replay performance against required recovery point and recovery time objectives.
• Sovereign control through an outage: Organizations choose the region, jurisdiction, and infrastructure where the recovery environment runs, while retaining authority over when it is promoted.
• A right-sized standby, without double-loading data: The recovery cluster can use a smaller footprint or cost-optimized cloud compute, avoiding the need to run every ingestion job twice into a duplicate production-scale warehouse.
• Integrated operations and enterprise support: Disaster Recovery is delivered within the supported EDB PG AI for WarehousePG stack, reducing add-on complexity and giving teams access to EDB's Postgres and distributed warehouse expertise.

Beyond Disaster Recovery, the same foundation gives enterprises a controlled path to migrate analytical workloads to the cloud and test peak-volume scenarios without disrupting production. Point-in-time restores can also offload intensive month- or quarter-end reporting, allowing teams to release the compute when the work is complete.

Open, sovereign analytics for the agentic enterprise

WarehousePG is an open source, Postgres-based MPP data warehouse designed for petabyte-scale analytics and high-concurrency workloads.

EDB PG AI adds the capabilities required to operate that open foundation in production, including unified observability, built-in high availability, real-time ingestion, lakehouse interoperability, in-database AI and machine learning, vector search, and 24x7 enterprise support. It can run on-premises, in a private VPC, or across clouds, giving enterprises control over data residency, infrastructure, and long-term economics.

Disaster Recovery adds another mission-critical control to that foundation: a supported path to protect analytical data and restore operations across infrastructure boundaries that also preserves the deployment flexibility and sovereignty that enterprises chose EDB PG AI to achieve.

Availability

Disaster Recovery for WarehousePG is generally available on EDB Postgres AI today. For more information, visit www.enterprisedb.com/products/warehousepg and read the announcement blog. 

About EDB

EDB Postgres^® AI (EDB PG AI) is the sovereign data and AI platform for the agentic enterprise. Built on Postgres, the world's leading open source database, EDB PG AI unifies transactional, analytical, and AI workloads in a single architecture, eliminating the data movement, ETL, and operational fragmentation that slow enterprises down. With governance enforced at the data layer and the flexibility to deploy on-premises, in hybrid environments, or across clouds, enterprises operationalize their data and AI on infrastructure they own and control, reaching production-ready sovereign AI in weeks, not months. As one of the most active contributors to the PostgreSQL project, EDB is deeply invested in the vitality of the global open source community. To learn more, visit www.enterprisedb.com.

Media Contact

Anwar Ahmad
[email protected]
347-206-5651

EnterpriseDB and EDB are registered trademarks of EnterpriseDB Corporation. Postgres and PostgreSQL are registered trademarks of the PostgreSQL Community Association of Canada and used with their permission. All other trademarks are owned by their respective owners.

SOURCE EnterpriseDB