You see this while chkdsk runs — sometimes hundreds of “File record segment 12345 is unreadable” lines — and it points to damage in the master file table, the index NTFS uses to find your files. It usually means bad sectors on a drive that’s starting to fail. Here’s what it means, what not to do, and how the data is recovered.
The master file table is where NTFS records every file. When chkdsk hits sectors it can’t read in that table, it reports them one by one. A flood of these usually means the drive has physical bad sectors and is failing — which is why imaging comes before any repair.
chkdsk /f or /r on a drive throwing these errors. Each pass forces the failing disk to re-read and remap the same bad sectors, stressing it further and potentially overwriting data that’s still recoverable. If the files matter, image the drive first and recover from the copy.NTFS keeps a master file table (the MFT): a catalogue split into 1 KB “file record segments,” one for each file, recording where it begins and ends on the disk. “File record segment N is unreadable” means chkdsk tried to read segment N of that table and the drive couldn’t return it. The file data itself is usually still on the platters; what can’t be read is the catalogue entry pointing to it — and most often that’s because the sector holding it has gone bad.
A handful of unreadable segments can come from a single bad patch, or a power cut caught mid-write. But when chkdsk lists them in volume — dozens or hundreds — it almost always means the drive has developed widespread bad sectors and is physically failing. The pattern to watch for: the scan crawls, the drive is slow, clicking or ticking, and the same errors return every time you run it. That’s a hardware fault, not something a software repair can put right.
Windows will keep offering to run chkdsk, and it’s tempting to let it finish. On a healthy drive with minor table damage it can rebuild the MFT. On a failing one it does harm: every /f or /r pass forces the disk to hammer the same unreadable sectors, accelerating the failure, and chkdsk can move or overwrite data as it “repairs” — turning a recoverable drive into a lost one. If you don’t know the drive is sound, stop. Don’t format it, and don’t run repeated scans.
The safe route is to clone the drive to a healthy one with a hardware imager that reads around bad sectors — retrying, slowing down, and skipping what won’t come back — so the original is never written to and never worked harder than it can bear. From that image the master file table is rebuilt, or where it’s too damaged the files are carved out by signature. A single hard drive or laptop or PC drive is from £300 +VAT; an external drive is sent whole; a USB stick or memory card is from £250 +VAT. Every job opens with a free diagnostic and a fixed written quote — most, no fix, no fee.
It’s a chkdsk message saying it couldn’t read part of the master file table (MFT), the index NTFS uses to find your files. The table is split into “file record segments,” one per file; an unreadable one usually sits in a bad sector. Your files are normally still on the disk — it’s the catalogue entry that can’t be read, often because the drive is starting to fail.
Usually not. The file data is typically still on the platters; what’s damaged is the table indexing it. As long as the drive is imaged before it deteriorates further, the data can usually be recovered by rebuilding the MFT from the copy or carving files out directly. The real risk is running repairs on a failing drive that overwrite recoverable data.
Only if you know the drive is healthy. A few unreadable segments on a sound disk can be repaired by chkdsk. But a long list of them usually means physical bad sectors and a failing drive — and every chkdsk pass then stresses the disk further and can overwrite data while trying to rebuild the table. If the files matter and you’re unsure, image the drive first.
Because the drive has developed widespread bad sectors. Each unreadable segment is a piece of the file table sitting in a sector the disk can no longer read, and a flood of them means the failure is spreading — typically the scan is very slow and the drive may be clicking. That’s a hardware fault; the fix is to image the drive read-only, not to keep scanning.
In most cases, yes. The drive is cloned with a hardware imager that reads around the bad sectors without stressing it, then the file system is rebuilt from that image. A single drive is from £300 +VAT, always confirmed in a written quote after a free diagnostic, and most jobs are no fix, no fee. Send the bare drive rather than the whole computer — a local shop can remove it in minutes if needed.
A flood of unreadable file record segments usually means a failing drive. Before you run chkdsk again, send it in and we’ll image it read-only past the bad sectors and rebuild the files from the copy — free diagnostic, fixed quote before any work.