Linux Pocket Guide

Name

md5sum files | --check file — coreutils

Synopsis

/usr/bin stdin stdout - file -- opt --help --version

The md5sum command prints a 32-byte checksum of the given files, using the MD5 algorithm (see http://www.faqs.org/rfcs/rfc1321.html for the technical details):

$ md5sum myfile
dd63602df1cceb57966d085524c3980f  myfile

Two different files are highly unlikely to have the same MD5 checksum, so comparing checksums is a reasonably reliable way to detect if two files differ:

$ md5sum myfile1 > sum1
$ md5sum myfile2 > sum2
$ diff -q sum1 sum2
Files sum1 and sum2 differ

or if a set of files has changed, using --check:

$ md5sum file1 file2 file3 > mysum
$ md5sum --check mysum
file1: OK
file2: OK
file3: OK
$ echo "new data" > file2
$ md5sum --check mysum
file1: OK
file2: FAILED
file3: OK
md5sum: WARNING: 1 of 3 computed checksums did NOT match

Two other programs similar to md5sum are sum and cksum, which use different algorithms to compute their checksums. sum is compatible with other Unix systems, specifically BSD Unix (the default) or System V Unix (-s option), and cksum produces a CRC checksum:

$ sum myfile
12410     3
$ sum -s myfile
47909 6 myfile
$ cksum myfile
1204834076 2863 myfile

The first integer is a checksum and the second is a block count. But as you can see, these checksums are small numbers and therefore unreliable, since files could have identical checksums by coincidence. md5sum is by far the best.