FSCK(8) System Manager's Manual FSCK(8)
fsck - file system consistency check and interactive repair
fsck -p[ # ] [ filesystem ... ]
fsck [ -y ] [ -n ] [ -sX ] [ -SX ] [ -t filename ] [ filesystem ] ...
The first form of fsck preens a standard set of file systems or the specified file systems. It is normally used in the script /etc/rc dur-
ing automatic reboot. In this case fsck reads the table /etc/fstab to determine which file systems to check. It uses the information
there to inspect groups of disks in parallel taking maximum advantage of i/o overlap to check the file systems as quickly as possible.
Normally, the root file system will be checked on pass 1, other ``root'' (``a'' partition) file systems on pass 2, other small file systems
on separate passes (e.g. the ``d'' file systems on pass 3 and the ``e'' file systems on pass 4), and finally the large user file systems on
the last pass, e.g. pass 5. Only one file system from each disk should be checked in each pass. A pass number of 0 in fstab causes a disk
to not be checked; similarly partitions which are not shown as to be mounted ``rw'' or ``ro'' are not checked. A number or range of num-
bers may be given after the -p to start or end the check at a specified pass number. A single number specifies the starting pass (e.g. -p2
or -p2-); a range specifies the starting and endig passes (-p2-4). A missing start means to start with pass 1 (-p-2). These can be used
to stop fsck to mount a temporary file system before continuing the check on larger file systems requiring scratch files.
The system normally takes care that only a restricted class of innocuous inconsistencies can happen unless hardware or software failures
intervene. These are limited to the following:
Link counts in inodes too large
Missing blocks in the free list
Blocks in the free list also in files
Counts in the superblock wrong
Allocated inodes in the free inode list
These are the only inconsistencies which fsck with the -p option will correct; if it encounters other inconsistencies, it exits with an
abnormal return status and an automatic reboot will then fail. For each corrected inconsistency one or more lines will be printed identi-
fying the file system on which the correction will take place, and the nature of the correction. After successfully correcting a file sys-
tem, fsck will print the number of files on that file system and the number of used and free blocks. Warning: kernel changes are required
to limit the types of inconsistencies, and fsck -p makes assumptions about the kernel's actions in repairing these. Vmunix, PDP-11 2.9BSD
with the UCB_FSFIX option and PDP-11 2.10BSD Unix kernels have the appropriate modifications; the -p option should not be used on other
Without the -p option, fsck audits and interactively repairs inconsistent conditions for file systems. If the file system is inconsistent
the operator is prompted for concurrence before each correction is attempted. It should be noted that a number of the corrective actions
which are not fixable under the -p option will result in some loss of data. The amount and severity of data lost may be determined from
the diagnostic output. The default action for each consistency correction is to wait for the operator to respond yes or no. If the opera-
tor does not have write permission fsck will default to a -n action.
Fsck has more consistency checks than its predecessors check, dcheck, fcheck, and icheck combined.
The following flags are interpreted by fsck.
-y Assume a yes response to all questions asked by fsck; this should be used with great caution as this is a free license to continue
after essentially unlimited trouble has been encountered.
-n Assume a no response to all questions asked by fsck; do not open the file system for writing.
-sX Ignore the actual free list and (unconditionally) reconstruct a new one by rewriting the superblock of the file system. The file
system should be unmounted while this is done; if this is not possible, care should be taken that the system is quiescent and that it
is rebooted immediately afterwards. This precaution is necessary so that the old, bad, in-core copy of the superblock will not con-
tinue to be used, or written on the file system.
The -sX option allows for creating an optimal free list organization. The following forms of X are supported for the following
-s4 (RP04, RP05, RP06)
-sBlocks-per-cylinder:Blocks-to-skip (for anything else)
If X is not given, the values used when the file system was created are used. If these values were not specified, then the value
400:9 is used.
-SX Conditionally reconstruct the free list. This option is like -sX above except that the free list is rebuilt only if there were no
discrepancies discovered in the file system. Using -S will force a no response to all questions asked by fsck. This option is useful
for forcing free list reorganization on uncontaminated file systems.
-t If fsck cannot obtain enough memory to keep its tables, it uses a scratch file. If the -t option is specified, the file named in the
next argument is used as the scratch file, if needed. Without the -t flag, fsck will prompt the operator for the name of the scratch
file. The file chosen should not be on the file system being checked, and if it is not a special file or did not already exist, it is
removed when fsck completes.
If no file systems are given to fsck then a default list of file systems is read from the file /etc/fstab.
Inconsistencies checked are as follows:
1. Blocks claimed by more than one inode or the free list.
2. Blocks claimed by an inode or the free list outside the range of the file system.
3. Incorrect link counts.
4. Size checks:
Directory size not 16-byte aligned.
5. Bad inode format.
6. Blocks not accounted for anywhere.
7. Directory checks:
File pointing to unallocated inode.
Inode number out of range.
8. Super Block checks:
More than 65536 inodes.
More blocks for inodes than there are in the file system.
9. Bad free block list format.
10. Total free block and/or free inode count incorrect.
11. Allocated inodes on the free inode list in the superblock.
Orphaned files and directories (allocated but unreferenced) are, with the operator's concurrence, reconnected by placing them in the
lost+found directory. The name assigned is the inode number. The only restriction is that the directory lost+found must preexist in the
root of the file system being checked and must have empty slots in which entries can be made. This can be accomplished manually by making
lost+found, copying a number of files to the directory, and then removing them (before fsck is executed). Mkfs(8) will automatically cre-
ate a lost+found directory.
Checking the raw device is almost always faster. The root device should not be checked using the raw device, however, since it cannot be
/etc/fstab default list of file systems to check
The diagnostics produced by fsck are intended to be self-explanatory. The exit codes with the -p option are 0 (no problems that weren't
fixed), 4 (root file system was modified), 8 (problems that couldn't be fixed) and 12 (fsck was interrupted).
filsys(5), fstab(5), crash(8), mkfs(8), mklost+found(8), reboot(8)
T. J. Kowalski, FSCK - The UNIX File System Check Program
Inode numbers for . and .. in each directory should be checked for validity.
3rd Berkeley Distribution FSCK(8)