Data Recovery Case File · Desktop Externals & Aging Drives · Busy Doing Nothing
The Disk Check That Hung at 99 Per Cent
His enquiry was one of the most carefully observed in this archive and contains its own diagnosis twice over. A 3TB Seagate Barracuda: "the failure occurred when copying a set of video files to the disk. The Windows disk checking function ran and hung permanently on 99% complete. Windows boots normally when the disk is removed, and Task Manager shows that when the disk is connected it is 100% in use, but only reading or writing for short periods every few seconds. This also crashes the PC." Two observations, and they describe the same underlying fact from different angles. A progress bar stuck at 99% and a drive at 100% activity with almost no throughput both mean the same thing: the drive is spending its time waiting rather than working.
| Media | 3TB Seagate Barracuda hard drive — failure occurring during a large write; filesystem check stalled indefinitely near completion; host reporting maximum device activity with minimal actual transfer |
| Reported situation | Failure during a bulk copy · disk check run and hung at 99% · host booting normally without the drive · device showing continuous activity with intermittent brief transfers · host crashes with the drive connected |
| Fault class | Read failures with escalating retries — activity time consumed by waiting; repair utility writing to a degrading surface |
| Equipment used | Atola Insight Forensic surface and error-rate assessment · PC-3000 Express with Data Extractor imaging under per-sector timeouts · healthy regions banked first, weak neighbourhoods deferred · losses reported per file |
The decode: what 99% means, and what Task Manager is measuring
Why the disk check hung near the end: a filesystem check works through the disk in phases, and the percentage reflects progress through the checking, not the health of what it finds. Hanging at 99% does not mean it nearly succeeded — it means it reached a region it cannot read and is now retrying indefinitely, because that is what these tools do when a sector will not answer. They have no facility to give up and move on. The number is meaningless; the stall is the finding.
Why running it was the risk: and this is worth saying although it is done. A disk check does not merely inspect — it writes, repairing structures it considers inconsistent, marking sectors bad and relocating what it can. On a drive that cannot reliably read, that is a sustained write-and-retry session on exactly the weakest ground, and it converts recoverable inconsistency into permanent structural loss. The tool most people reach for first is the one that most often makes these worse.
What Task Manager is actually telling him: the sharpest observation in the enquiry. The "active time" figure does not measure throughput — it measures the proportion of time the device has requests outstanding. A healthy drive answers in milliseconds, so it is rarely at 100%. A drive at 100% active with brief bursts of transfer every few seconds is one where nearly all the time is spent waiting on reads that eventually complete after retries, or never complete at all. So his two observations agree: the drive is busy waiting, not busy working.
Why it crashes the machine: the same fault seen from above. The storage driver blocks waiting for a device that does not answer, and because storage handling is shared machinery deep in the operating system, the system eventually halts rather than continue in an unknown state. That the machine boots normally with the drive removed confirms the fault belongs entirely to the drive.
Why the failure happened during a copy: bulk writing is the most demanding thing most people ask of a drive — sustained, continuous, and touching regions that may not have been used in years. A drive with marginal areas will cope with everyday use and fail under exactly that load. The copy did not break a healthy drive; it asked more of it than anything previously had.
On the bench
The drive was not connected to a computer again, since three separate symptoms had already established what that produces. The Atola Insight Forensic assessed surface condition and error rates directly, converting his observations into a measured map of where the drive still read cleanly. Imaging ran on the PC-3000 Express under Data Extractor with per-sector timeouts enforced — the decisive difference from both the disk check and the copy that failed: a failing sector is given a strict budget and deferred rather than retried indefinitely. The healthy expanse was captured first at full speed, weak neighbourhoods revisited on later passes, and every unreadable region resolved to the files it affected.
The outcome
The drive imaged under timeout control with the weak regions re-read, and the contents delivered with losses named per file. 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, for anyone whose disk check has stalled: a percentage measures progress through the checking rather than success, so hanging near the end means the tool has met a region it cannot read and is retrying forever — the number is meaningless and the stall is the finding; and a device reported at 100% activity with almost no transfer is measuring time spent waiting, not working. Both describe a drive failing reads. Stop the check, disconnect the drive, and do not run it again.
Disk check stuck near the end
Stop it and unplug the drive. A progress percentage measures how far through the checking the tool has got, not how close it is to succeeding — so hanging at 99% means it has reached a region it cannot read and is retrying indefinitely, because these tools have no way to give up and move on. The number tells you nothing; the stall is the diagnosis. Don't run it again, and don't let anything else attempt a repair, because a disk check writes to the drive: it repairs structures, marks sectors and relocates data, which on a failing surface is a long write-and-retry session on exactly the weakest ground. If your task manager shows the drive at 100% active with almost no throughput, that's the same finding from another angle — active time measures how long requests are outstanding, so a drive that's permanently busy and barely transferring is waiting rather than working.
Both mean the same thing — call Glasgow Data Recovery on 0141 404 0294; surface measured, imaged healthy-regions-first under per-sector timeouts, weak areas re-read patiently.
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.