Data Recovery Case File · Desktop Externals & Aging Drives · Specific Folders, Specific Regions
Dropping Off Is the Host Giving Up Waiting
His enquiry reports two behaviours that describe one fault. A secondary drive used for documents and archives: "it has become difficult to access and likes to unmount itself, and certain folders cannot be accessed because it says that" they are unavailable. A drive that unmounts itself has not been disconnected — the host asked it for something, waited longer than it was prepared to, and removed the device. And the fact that only certain folders fail is the most useful thing in the message.
| Media | 2TB internal hard drive used as secondary document and archive storage — access degrading; volume dismounting during use; specific directories inaccessible |
| Reported situation | Drive serving as secondary storage for documents and archives · access becoming difficult · volume dismounting itself during use · specific folders reporting as inaccessible while others open normally · contents required |
| Fault class | Localised progressive read failure — affected regions holding specific directory entries; substantial majority of the medium still reading |
| Equipment used | Drive removed from use immediately · Atola Insight Forensic error-rate assessment establishing failure distribution · imaged write-blocked under strict per-sector timeouts with retries capped · healthy expanse captured first · losses mapped per file |
The decode: why it drops, and what selective failure tells you
Why a drive appears to disconnect itself: the operating system issues a read and waits for an answer. It waits generously, but not indefinitely — and when a request exceeds its patience it concludes the device has gone and removes it from the system. Nothing was unplugged and no connection failed. The drive simply did not reply within the time allowed.
What causes the delay: retries. A drive that cannot read a region does not fail immediately — it repositions, tries again, and escalates through an internal recovery sequence taking seconds per sector. Enough of those in one operation and the host's timeout expires, which is why the drive vanishes mid-use rather than at connection.
Now the diagnostic part, and it is genuinely good news. Only certain folders fail. If the drive were failing generally, everything would be equally difficult — so selective failure means the unreadable regions are specific ones, and they happen to hold the directory entries for those particular folders. Everything else reads normally because the areas holding it are intact.
Why that matters for expectations: localised damage means the great majority of the medium is still good. This is not a drive that has generally deteriorated; it is a drive with defined bad regions and a large healthy remainder — which is the difference between capturing most of the content and capturing very little.
Why acting now matters, and it is the urgent part: "has become difficult" describes a change over time, which means the process is ongoing. The regions failing today are more than were failing last month. Every browse touches the failing areas, because that is what browsing does — it reads directory entries — and each attempt is more sustained work on exactly the wrong part of the drive.
What must not be run: a consistency check. It will read the whole volume, which on a drive with defined bad regions means hours of retries at maximum effort, and it repairs by discarding entries it cannot verify — which are precisely the entries in the failing regions.
Why the capture approach inverts the drive's own behaviour: imaging under strict per-sector timeouts caps retries deliberately, takes the healthy expanse quickly, and returns to weak regions on later passes rather than letting one bad area consume the session.
On the bench
The drive was removed from use immediately, browsing reading directory entries and therefore touching the failing regions directly. The Atola Insight Forensic assessed error-rate distribution to establish failure localisation — selective folder failure indicating defined bad regions holding specific entries rather than general deterioration, which materially changes what can be captured. Imaging ran write-blocked under strict per-sector timeouts with retries capped, the healthy expanse captured first, and losses resolved to individual files.
The outcome
The failure distribution established first, the healthy expanse captured quickly and weak regions revisited on later passes. 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: your drive is not disconnecting — the system is removing it after a read takes longer than it will wait, which happens because the drive is retrying. And only certain folders failing means the bad regions are defined ones holding those entries, so most of the medium is still good.
Drive that drops off and refuses particular folders
Stop browsing it — that's the activity most likely to make things worse, because browsing reads directory entries and those are exactly what's failing. Nothing is disconnecting: the system issues a read, waits as long as it's prepared to, and removes the device when the answer doesn't arrive. The delay is the drive retrying, repositioning and escalating through an internal recovery sequence that takes seconds per sector. The encouraging part is that only certain folders fail, which means the unreadable regions are specific ones rather than general deterioration — so most of the drive is still good. Don't run a consistency check; it reads everything and discards what it can't verify.
Stop browsing it — call Guildford Data Recovery on 01483 901310; failure distribution established first, imaged under capped per-sector timeouts, healthy expanse captured before weak regions are revisited.
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.