← Back to Blog

DAOS Version 2.6 Release: Enhanced Performance and Reliability

daos@storage-node-01 ~
$ daos version daos version 2.6.0 $ dmg pool create --size 50GB tank Creating DAOS pool with automatic storage allocation: 50 GB total Pool created with 6.00% SCM/NVMe ratio UUID : 8a05bf3a-a088-4a77-bb9f-df989fce7cc8 Service Ranks : [1-5] Storage Ranks : [0-15] Total Size : 50 GB Tier 0 (SCM) : 3.0 GB (188 MB / rank) Tier 1 (NVMe) : 47 GB (3.0 GB / rank) $ daos pool query tank Pool 8a05bf3a, ntarget=32, disabled=0, leader=0, version=1 SCM Total: 3.0 GB Free: 2.8 GB NVMe Total: 47 GB Free: 46 GB Rebuild idle, 0 objs, 0 recs

DAOS 2.6 is out. The headline: small file I/O is about 30% faster, memory overhead is down 15%, and there's a meaningful batch of reliability and usability work underneath.

What changed

On the performance side, the biggest win is small file workloads. Operations that previously bottlenecked on metadata path overhead now run roughly 30% faster. Memory utilisation is tighter across the board, which matters when you're running dense multi-node deployments where every gigabyte of DRAM counts.

Reliability got attention too. Error detection and recovery mechanisms are more robust, data integrity checks run more efficiently, and fault tolerance for hardware failures has been hardened. These aren't flashy changes, but they're the kind of work that keeps production clusters running without surprises.

On the usability front, cluster configuration is simpler, monitoring and diagnostics tools have been improved, and integration with existing storage ecosystems is smoother. The goal has always been to make DAOS easier to operate at scale, and 2.6 pushes that forward.

Getting started

If you're looking to upgrade, check the DAOS documentation or reach out to us via the contact form and we'll help you through it.