---
title: "OrioleDB public beta: a Postgres storage engine for update-heavy tables"
description: "OrioleDB, an open-source storage engine for Postgres, is now in public beta on every Supabase plan. It replaces the heap with an undo log, 64-bit transaction IDs, index-organized tables, and lock-free page reads to eliminate dead tuples, reduce bloat, and improve concurrency. Performance benchmarks on DBArena show that when data exceeds memory, OrioleDB achieved up to 1.8x more transactions per minute than heap storage—113,850 on an 8XL instance—and about 46% more throughput per dollar at small sizes. The engine is chosen at project creation, billed the same as standard Postgres, but has limitations such as no Point-in-Time Recovery, unsupported VACUUM FULL, and experimental non-B-tree indexes. The Supabase team recommends benchmarking your own workload, as OrioleDB helps most on update-heavy tables with primary-key lookups, data larger than memory, and high concurrency."
---

# OrioleDB public beta: a Postgres storage engine for update-heavy tables

[Supabase](https://yomu.fyi/company/supabase) · aditya\_maruvada,daniel\_mitterdorfer · Oct 28, 2026

**Type:** Announcement

## Summary

OrioleDB, an open-source storage engine for Postgres, is now in public beta on every Supabase plan. It replaces the heap with an undo log, 64-bit transaction IDs, index-organized tables, and lock-free page reads to eliminate dead tuples, reduce bloat, and improve concurrency. Performance benchmarks on DBArena show that when data exceeds memory, OrioleDB achieved up to 1.8x more transactions per minute than heap storage—113,850 on an 8XL instance—and about 46% more throughput per dollar at small sizes. The engine is chosen at project creation, billed the same as standard Postgres, but has limitations such as no Point-in-Time Recovery, unsupported VACUUM FULL, and experimental non-B-tree indexes. The Supabase team recommends benchmarking your own workload, as OrioleDB helps most on update-heavy tables with primary-key lookups, data larger than memory, and high concurrency.

## Context

Postgres’s default heap storage accumulates dead tuples on updates and deletes, requiring VACUUM cleanup. On write-heavy tables, autovacuum can fall behind, causing bloat, slower queries, and management overhead. At high concurrency, the shared buffer pool and checkpoint I/O create further bottlenecks.

## Approach / What changed

OrioleDB replaces the heap via Postgres’s Table Access Method. It uses an undo log instead of dead tuples, 64-bit transaction IDs, index-organized tables, no buffer mapping, and copy-on-write checkpoints to reduce bloat, improve concurrency, and eliminate VACUUM overhead.

## Takeaways

- OrioleDB uses an undo log and 64-bit transaction IDs to eliminate dead tuples and the need for VACUUM on update-heavy tables.
- Performance improvements are largest when the data set does not fit in memory; on a Small instance, OrioleDB delivered about 46% more throughput per dollar than heap storage.
- OrioleDB must be selected at project creation and cannot be retrofitted to existing projects; it also lacks Point-in-Time Recovery and support for VACUUM FULL.

## FAQ

### How does OrioleDB avoid dead tuple accumulation?

OrioleDB writes previous row versions to an undo log instead of leaving dead tuples in the heap. The undo log is reclaimed automatically, so tables accumulate no dead tuples and VACUUM has nothing to clean.

### What limitations should I know before using OrioleDB in beta?

The beta cannot be added to existing projects, lacks Point-in-Time Recovery, and does not support VACUUM FULL or CLUSTER. Non-B-tree indexes like pgvector HNSW run through an experimental bridge, and certain extensions like plv8 and timescaledb-apache are not included.

### Under what workloads does OrioleDB offer the biggest performance gain?

OrioleDB helps most on tables with frequent updates or deletes, when the data set is larger than memory, on large instances with many concurrent connections, and when most queries use primary-key lookups. Read-mostly workloads fitting in memory see smaller differences.

**Tags:** [Benchmark](https://yomu.fyi/topic/benchmark), [OrioleDB](https://yomu.fyi/topic/orioledb), [Performance](https://yomu.fyi/topic/performance), [Postgres](https://yomu.fyi/topic/postgres), [Storage](https://yomu.fyi/topic/storage)

- Source: [Supabase](https://supabase.com/blog/orioledb-public-beta)
- Source URL: https://supabase.com/blog/orioledb-public-beta
- Ingested by Yomu: 2026-10-28T18:01:11.677Z

[Read original post](https://supabase.com/blog/orioledb-public-beta)
