ARCHIVED: What is a lost cluster?

A cluster is a unit of storage space on an MS-DOS/Windows-formatted hard drive (in Unix operating systems, the concept is called a block). A lost cluster (in Unix, a lost block or orphan) is one that is not used by a file, but that is not listed as free in the file allocation table (FAT) in Windows 9x, the Master File Table (MFT) in NT-based operating systems like Windows 2000, XP, or 2003, or the Superblock in Unix and Linux operating systems. Lost clusters can sometimes contain parts of corrupted files; however, they are usually blank space that was originally allocated to a program and for some reason has not been released. Lost clusters can occur when there is a loss of power to your computer or when programs are not shut down properly, so that data waiting in the disk cache to be written to the hard drive is lost; this happens even though the FAT, MFT, or Superblock has been updated to say that it was actually written. Clusters that are really empty but are listed in the FAT, MFT, or Superblock as containing part of the file are lost clusters.

In Windows, you can attempt to reclaim lost clusters by running ScanDisk or CHKDSK (in Windows 2000 and later). These programs collect the data in lost clusters and put it into files named file####.chk. Most of the time these files are useless, and you can delete them. If you want to view these files, however, you can open them in a word processor such as Microsoft Word. You may find some data in the file fragments that might be useful. Usually, though, you will want to just delete these files to reclaim the space.

In Unix, you can use the program fsck (Filesystem Check) to try to fix orphans. The data from those orphaned blocks is put into a lost+found directory, and files are named after the inode where the blocks were discovered.