When shared storage fails, all the servers that were using it go down as well. We rebuild SAN estates (LUNs offline, damaged storage pools, VMFS datastores that won't mount) for both businesses and IT providers in Swindon and Wiltshire, from a single shelf right up to multi-shelf arrays. Enterprise server data recovery is what we typically do day to day on the bench.
All SAN jobs come with a free diagnostic. You get a single fixed quote in writing before anything is started.
In most cases you only pay if we recover your data. Electronic and mechanical failures and chip-level work are 50% upfront, with the balance only if we recover your data, while BitLocker, ransomware, DVR and forensic jobs are paid upfront once you accept the fixed quote. You'll find full pricing on the data recovery cost page.
We begin each SAN job by matching the symptoms to the fault, and the twenty listed here cover almost everything that makes it to our bench.
Enterprise shelves tend to lose members in groups. Each failed member is repaired and imaged on its own before we work through the maths of the group.
The block device your servers ran on disappears, so we start rebuilding from the pool metadata and the RAID groups and work upwards.
When the mapping of blocks to LUNs breaks down, we recreate it from the raw on-disk structures that it originally described.
When the firmware dies mid-update it takes the management plane down with it, but the data behind it never actually moved.
Redundant pairs fail together more often than most brochures would have you believe, and when they do, recovery goes directly to the drives.
When the VMware layer collapses over healthy LUNs, we extract the datastores and their VMDKs and then repair them.
Provisioning errors take only seconds to make, but what survives depends on the number of writes made after the error, so the platform should be frozen as soon as possible.
When one link in the chain breaks, all the volumes that depend on it fail as well, so we stitch the whole chain back together at capture level.
If a pool has advertised more storage space than it actually had, it will corrupt every thin volume that draws on it.
The combination of dead cache batteries and a power event leaves parity disagreeing with data across the whole group, and we reconcile this from images.
NetApp aggregates and Dell EMC pools can lose their LUN maps, so our rebuild starts at the RAID groups and works up through the layers.
Enterprise sector formats typically defeat desktop software tools, but the lab fabric images them natively.
HPE's small business SAN staple often comes in with damaged vDisk metadata, and rebuilding it is a commonly performed task in the lab.
After a power event, IBM V3700 and V5000 pools may lose some extents, and we remap those directly from the drives.
The box took the target definition with it, but the volume associated with that target generally survives and can be extracted.
If the dedupe index gets corrupted, every pointer breaks at once, so we rebuild it until the shared blocks resolve again.
In a botched failover both controllers wrote on their own, so the two halves that drifted apart are reconciled at image level.
Kit that is long past vendor support can fail with no formal recovery path left, but lab recovery can still go ahead whatever the contract status.
Replacement drives on different firmware can cause issues when they are placed into an older group. We stabilise and image them before the group is reconstructed.
Changes to the fabric or zoning can cause healthy LUNs to not be seen by any servers, so we rebuild the mapping and pass the volumes back.
A SAN is simply a stacked set of abstract layers. Physical drives make up RAID groups, those RAID groups make up pools, the pools are portioned out as LUNs, and file systems or hypervisor datastores sit on top. When recovering, you work down through that stack and then back up again. First image the drives, then bring the RAID groups back, rebuild the pool metadata, pull out the LUNs, and finally open the VMs and databases inside. We have experience working with Dell EMC, HPE MSA/EVA, NetApp, IBM and Fujitsu hardware, and we know how to work within NDAs and change control paperwork without any friction.
No business really wants raw blocks back. They want the three machines that run everything online again. So we mount the datastores from the rebuilt LUNs, take the VMDK and VHDX files out for repair and move named SQL or Exchange workloads to the front of the line, so operations can restart while the remaining terabytes are still copying in the background.
SAN work needs enterprise plumbing just as much as it needs recovery skill, and the lab keeps both on hand:
Enterprise storage shelves and drives connect over the lab's own Fibre Channel (FC), SAS and iSCSI plumbing, and that includes exotic 520/524/528 sector sizes.
Once any failed members have finished their time on the PC-3000 rigs, the RAID groups are captured side by side.
Vendor metadata is parsed and rebuilt in sequence: RAID groups, then pools, then LUNs, and finally the file systems on top.
Once the LUNs are rebuilt they host their datastores again, and each virtual disk is then lifted out and repaired in turn.
Named SQL Server and Exchange workloads are verified and moved to the front of the queue, so critical services restart first.
All original drives stay read-only from the moment they're booked in until the day they go home.
An unfamiliar badge is rarely unfamiliar hardware, because the enterprise market rebadges the same building blocks over and over. Pools that lose their LUN maps are rebuilt from the RAID groups upwards; a double controller failure simply sends the work straight to the drives; and 520/528-byte enterprise sectoring is imaged natively on the lab fabric instead of fighting desktop adapters.
Before anything comes off the rack, tag every drive with its shelf and slot position, because enterprise layouts forgive nothing where order is concerned. Get a quote first, by phone (0800 689 0668) or with the free online diagnostic. Only the tagged drives are posted (photograph the configuration first); the shelves and controllers stay on site.
Most customers get a quote first, by phone or with the free online diagnostic, and then post or courier their media to us.
Sending a drive from a computer, laptop, MacBook, iMac, CCTV / DVR or server? Please remove the internal hard drive or SSD and send us just the drive (from a server, every disk, each labelled with its bay), as we don't offer a drive-removal service. We don't recover storage soldered to a motherboard (for example T2 and Apple Silicon Macs and some thin laptops), only drives that can be removed and sent to us. If you're not sure which kind you have, ring us first.
↓ Print the booking-in & shipping form (PDF)
Mark the package for the attention of Oxford Data Recovery and we'll ring you as soon as we've diagnosed your media.
Not sure what to send? Ring 0800 689 0668 first or use the free online diagnostic.
Free diagnostic, fixed quote, and in most cases you only pay if we recover your data. Start now or call the freephone.