Object and Scale-Out Storage Data Recovery

Object and scale-out platforms distribute both data and metadata across many nodes, usually with erasure coding rather than RAID. There is no single array to examine: a bucket, file system or namespace only exists as an assembled view of fragments held on dozens of devices, and the mapping that assembles them is itself distributed.

Overview

These systems fail in a different order from traditional arrays. Individual drives and even whole nodes can be lost without data loss, but a metadata service that loses quorum, a database that becomes inconsistent, or an erasure-coded stripe that loses more fragments than its scheme tolerates can make an entire namespace unreadable while the media stays healthy.

Because most platforms in this category are software running on commodity servers, the recovery scope is defined by the cluster rather than by a chassis. Drives from every contributing node, the node identities, the pool or erasure-set topology and any separate metadata devices normally have to be preserved together.

The layer above matters as much as the layer below. Objects written by S3 clients carry keys and user metadata that must be reconstructed alongside the payload; parallel file systems used in HPC and AI workloads carry directory structures and striping information in dedicated metadata targets that are recovered before any file content can be resolved.

The most damaging post-failure actions here are administrative: re-initialising a cluster, recreating pools or erasure sets, re-imaging nodes, wiping and re-adding a node to force a rebuild, or letting a rebalance run over an already-degraded layout.

Platform Families We Evaluate

• Qumulo Core Data Recovery — Scale-out NAS with real-time file analytics — Qumulo Core on Qumulo/HPE/Supermicro nodes and Cloud Native Qumulo.
• VAST Data Platform Data Recovery — Disaggregated shared-everything (DASE) all-flash platform — CNodes and DBoxes with storage class memory and QLC flash.
• WEKA Data Platform Data Recovery — Distributed NVMe-native file system — WekaFS clusters with N+2/N+4 protection and tiering to object storage.
• Scality RING Data Recovery — Peer-to-peer distributed object storage — RING nodes, DHT-based keyspace, ARC erasure coding and replication classes of service.
• Cloudian HyperStore Data Recovery — S3-compatible object storage with Cassandra-backed metadata — replication and erasure-coding storage policies, object lock support.
• MinIO Enterprise Object Storage Data Recovery — S3-compatible distributed object storage — server pools, erasure sets, xl.meta object metadata, bare metal and Kubernetes deployments.
• DDN SFA & EXAScaler Data Recovery — HPC storage family — DDN SFA declustered-RAID block arrays paired with EXAScaler Lustre file systems (MDT/MGT/OST on ldiskfs or ZFS).

Related Services

• NAS Data Recovery
• Enterprise Data Recovery
• Unix & Linux Data Recovery
• RAID Data Recovery

What Not To Do Before an Evaluation

• Do not run rebuilds, reconstructions or re-initialisations against an array that has already lost more drives than its protection level allows.
• Do not recreate pools, aggregates, disk groups, storage pools or clusters — these operations write new metadata over the structures a recovery needs.
• Do not swap drives between slots, and do not reorder shelves. Record the original slot and shelf positions before removing anything.
• Do not run file-system repair tools against production volumes before the underlying storage layer has been evaluated.
• Do not restore a backup or replication set over the affected volumes until the recovery scope has been assessed.
• Do not eradicate deleted volumes or empty recycle/destroyed states on platforms that hold deleted data for a retention window.

Our Evaluation and Recovery Process

• Intake and platform identification — array model, generation, firmware, protection layout and the sequence of events that led to the failure.
• Read-only evaluation of the media and array structures, including assessment of drive health and the extent of any physical damage.
• Forensic imaging of all contributing media, with cleanroom work where drives require it. Originals are preserved unaltered.
• Reconstruction of the storage layer — pools, aggregates, parity groups, chunklets, extent groups or objects — from the images.
• Extraction of the layers above: file systems, virtual machines, databases, mailboxes and shares.
• Verification against a file list and customer-nominated critical data, followed by secure return on encrypted media.