On paper this Sheffield enquiry is a negotiation, but a technical decision sits inside it. There is "a SAN with 24 disks" holding "two RAID 5 arrays", a rival firm quoting a five-figure sum and promising to "do it all by close of play on Sunday", and one blunt question for us: can we "improve on their price or timeframe". The honest reply is that nobody should promise Sunday until the disks have been imaged, because with RAID 5 the schedule belongs to the weakest member, never to the calendar.
| Media | Enterprise SAN shelf, twenty-four disks configured as two twelve-member RAID 5 arrays, state of individual members unknown at enquiry |
| Reported situation | Recovery required across both arrays · a competing quotation in hand with a fixed weekend deadline · enquirer connected to datacentre operations with recurring requirements · price and timeframe both in negotiation |
| Fault class | Multi-member array recovery where the failure detail is not yet stated — member condition, not array count, determines difficulty, duration and cost |
| Equipment used | Every member imaged individually under hardware write-blockers before any assembly · DeepSpar Disk Imager for degraded members with per-head control · PC-3000 UDMA where a member is unstable or firmware-impaired · controller metadata and parity analysed in UFS Explorer RAID and R-Studio Technician on the images · XOR consistency and checksum verification across each stripe set before extraction |
RAID 5 survives one missing member per array and not two. Whatever took the arrays down, the recovery question is how many of the twenty-four disks read cleanly. Twelve-member RAID 5 is already the aggressive end of that geometry: one parity disk's worth of protection stretched across eleven neighbours.
The members are the same age, and that is not a comfort. Arrays are built from disks bought together, from the same production window, doing identical work for identical hours. When one fails from wear, its siblings are statistically close behind — which is exactly why the second failure so often arrives during the stress of a rebuild.
A fixed weekend deadline is only honest after the imaging, never before it. If all twenty-four members image cleanly, a weekend is realistic. If two members in one array are degraded, the pace is set by coaxing sectors off failing platters, and no promise made on Tuesday changes that physics.
Nothing is ever assembled on the original disks. Every member is imaged first, and the arrays are reassembled virtually from the images. A rebuild forced on original hardware is the single most common way a recoverable RAID case becomes an unrecoverable one, because it writes to every member at once.
Parity order and layout are read from the disks, not assumed. Controller metadata, stripe size, rotation and member order are established from the evidence and verified with XOR consistency checks before a single file is extracted — a mis-ordered assembly produces plausible-looking corruption.
On price, the honest structure is conditional. A single figure quoted before member assessment is either padded against the worst case or gambling on the best one. The defensible quotation prices the imaging, then fixes the recovery figure once the true condition of both arrays is known.
Each member was catalogued, serial-matched to its shelf position, and imaged individually under hardware write-blockers — healthy members directly, degraded members on DeepSpar Disk Imager with per-head control, and any unstable member handled through PC-3000 UDMA. Both arrays were then assembled virtually in UFS Explorer RAID with the layout confirmed against controller metadata, cross-checked in R-Studio Technician, and verified stripe-by-stripe with XOR consistency checks before extraction to new media.
Twenty-four members imaged individually, two arrays assembled virtually from images, parity-verified, and extracted. Free assessment, one fixed written figure including VAT; where a drive has to be opened, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode: the deadline belongs to the weakest disk. Image first, and the calendar becomes a promise you can keep instead of a risk you are taking with someone else's arrays.
Do not let anyone rebuild on the original disks, and treat a fixed completion date quoted before member-by-member imaging as a warning sign rather than a selling point. Power the shelf down, keep the members in order, and note which slots reported failure. The number that decides everything is how many disks image cleanly — establish that first, and price and timeframe follow from evidence.
Our case files are written up from genuine enquiries our lab has handled for customers across Sheffield and South Yorkshire, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery approach our engineers apply to that fault, using the equipment listed.
Free diagnostic, fixed quote, no fix no fee — start now or call the freephone.