This is chkdsk reporting that it could not read part of the master file table. A handful of segments is a repairable file system. Hundreds is a failing disk, and continuing to run chkdsk makes it worse.
Each segment is a 1 KB record describing one file. A few unreadable ones is localised damage. A flood means the sectors holding the table are failing across the disk.
NTFS stores its master file table as a series of 1 KB file record segments, one per file or folder, holding the name, timestamps, permissions and the list of clusters where the contents live. When chkdsk reports segment 12345 as unreadable, it means it asked the drive for that record and did not get valid data back.
The file’s contents are almost always still on the disk. What is missing is the entry describing where they are, which is why these situations are usually recoverable — and why the wrong repair destroys that possibility.
A handful, on a drive that is otherwise fast and quiet, generally means localised corruption — an interrupted write, a power loss, a bad shutdown. chkdsk will often repair that successfully and the drive will carry on for years.
Hundreds or thousands, with the scan crawling and the drive slow or audibly working hard, means the sectors holding the table cannot be read reliably. That is physical, it spreads, and it will not be fixed by software. The tell is consistency: run it again and a healthy drive reports the same short list, while a failing one reports a longer one.
chkdsk exists to produce a valid file system, and where it cannot read a record it achieves that by removing the entry or truncating the file. On a sound drive that is a fair trade. On a failing one it discards records that a recovery could have reconstructed from elsewhere on the disk, and the /r switch adds a full surface scan — hours of reading every sector on a disk already struggling to read the ones it needs.
Files that end up in found.000 as numbered .chk files are the visible result: contents preserved, names and structure gone.
If the drive still reads, copy the irreplaceable material off now, before running anything else. Then stop using it. Recovery means imaging read-only past the bad sectors with hardware built for the job, and rebuilding the master file table from that image — or carving files by signature where the table is beyond repair.
A single drive is from £300 +VAT after a free 48-hour diagnostic, most jobs no fix, no fee. Send the bare drive rather than the whole machine.
That chkdsk could not read a 1KB record in the NTFS master file table describing one file — its name, timestamps and the clusters holding its contents. The file’s data is usually still on the disk; the entry pointing at it could not be read.
It depends how many. A handful on a drive that is otherwise quick and quiet is usually localised corruption from a bad shutdown. Hundreds, with the scan crawling, means the sectors holding the table are failing — that is physical and it spreads.
Only if the count is small and the drive is healthy. chkdsk produces a consistent file system by discarding records it cannot read, and /r adds a full surface scan. On a failing drive that is both the wrong tool and hours of the wrong stress.
Fragments chkdsk recovered but could not reattach to a filename, saved as numbered files. The contents may be intact but the names and folder structure are gone — which is what happens when the file records themselves could not be read.
Usually. The drive is imaged read-only past the bad sectors and the master file table rebuilt from the copy, or files carved by signature where the table is beyond repair. A single drive is from £300 plus VAT after a free 48-hour diagnostic.