Data Recovery Case File · Mac & Apple Systems · The Rescue Was the Load
Backing Up a Failing Drive Asks It to Read Everything It Holds, at Once
Her enquiry describes a rescue attempt that consumed what it was trying to save. A machine running slowly with fans up, diagnosed as needing its system reinstalled, taken to a repair shop who "tried to do a full backup of my data as I had no previous backup — but it crashed, and now it shows a folder with a question mark." A whole-drive backup is the most demanding thing you can ask of a drive, and it was asked of one already in difficulty.
| Media | Internal drive of a desktop Mac — performance degradation and elevated cooling activity, followed by a failed whole-volume backup attempt and subsequent inability to locate a startup system; approximately 20,000 photographs held |
| Reported situation | Machine running slowly with continuous cooling activity but no excess heat · troubleshooting concluding the system required reinstallation · machine taken to a repair shop · whole-volume backup attempted with no prior backup held · backup crashing partway · machine subsequently unable to locate a startup system · approximately 20,000 photographs required |
| Fault class | Progressive read failure revealed and advanced by a whole-volume read — early symptoms consistent with retry latency; startup structures affected after the failed attempt |
| Equipment used | Early symptoms reinterpreted as read latency rather than software performance · whole-volume backup identified as sustained full-surface load · drive removed and imaged write-blocked under strict per-sector timeouts · healthy regions captured first with marginal areas revisited across passes · photographic content carved where structures were unrecoverable |
The decode: what the early symptoms actually were, and what the backup did
What slowness with active cooling and no heat usually means: the machine is busy waiting rather than working. Fans respond to load, and a system repeatedly retrying reads keeps processes active without generating the heat that real computation would — which is a recognised signature of a drive answering slowly.
Why that matters in hindsight: the symptom was being read as a software problem, and the recommended remedy was reinstalling the system. Reinstalling addresses software on the assumption that the storage beneath it works — so the diagnosis pointed at the wrong layer from the start.
Why the backup was nevertheless the right instinct: she had no copy, and getting one before any reinstallation is exactly the correct order. The shop did the sensible thing in the sensible sequence — the difficulty is what a whole-volume backup requires of the drive it reads.
What that requirement is: reading every occupied region, continuously, for as long as the job takes. Ordinary use touches a small area intermittently and leaves the drive idle most of the time, whereas a full backup asks the entire surface to perform without pause.
Why a drive with developing read failure fails during exactly that: the regions it struggles with are reached. Areas that ordinary use never visits are visited, and each difficult one absorbs retry cycles until the drive can no longer sustain the operation.
What the crash then did: left the operation incomplete and the drive in a worse state than it started. The startup structures failing afterwards is consistent with regions that were marginal before becoming unreadable during the sustained load.
Why this is not a criticism of the shop: the alternative — reinstalling with no copy — would have been considerably worse. What was missing was an image rather than a file-level backup, which is a distinction most general repairers do not draw.
Why an image is the right tool for a suspect drive: a backup reads files in folder order and stalls at the first unreadable one. An image proceeds by condition, capping the time spent on any region and taking healthy areas first — so it captures the maximum while the drive is at its best.
Why the photographs remain a reasonable prospect: the startup structures are a tiny fraction of the volume. Their loss prevents the machine starting and says nothing about twenty thousand image files, which sit elsewhere and are found by signature even where the structure cannot be rebuilt.
What must not happen now: no further backup attempts, no reinstallation, and no repair utilities. The drive comes out of the machine and is imaged before anything else is tried.
On the bench
Early symptoms were reinterpreted as read latency rather than software performance — slowness with elevated cooling activity and no corresponding heat indicating processes waiting on retried reads. The whole-volume backup was identified as sustained full-surface load, reaching regions ordinary intermittent use never visits and absorbing retry cycles at each, whereas an image caps time per region and takes healthy areas first. Imaging ran write-blocked under strict per-sector timeouts with marginal areas revisited across passes.
The outcome
Early symptoms reinterpreted as read latency, the backup identified as full-surface load, and the drive imaged under capped timeouts with healthy regions first. 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: a whole-volume backup asks the drive to read everything it holds without pause, which is far more than daily use. Your startup structures are a tiny fraction of the volume — their loss says nothing about the photographs.
Backing up a drive you already suspect
Ask for an image rather than a file-level backup, since the distinction matters most on a drive that may be failing. A backup reads files in folder order and stalls at the first unreadable one, potentially spending hours there and never reaching healthy material beyond it. An image proceeds by condition instead, capping the time spent on any region and capturing the good areas first while the drive is at its best. Read the early warning signs too: slowness with fans running but no real heat often means processes waiting on retried reads rather than a software problem, which is why reinstalling addresses the wrong layer.
Stop before trying again — call Easy Data Recovery on 028 9002 0144; early symptoms reinterpreted as read latency, imaged under capped timeouts with healthy regions captured first.
Request a quote online →
Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.