[ILUG] Request to ILUG gurus: comparing files on different machines

[Brian Foster]

  | Date: Fri, 06 Oct 2006 12:18:17 +0100
  | From: Thomas Pedoussaut <thomas at pedoussaut.com>
  | 
  | Timothy Murphy wrote:
  | > I want to determine which files in directory foo/ on remote machine
  | > A are duplicates of files in directory foo/ on my current machine.
  | > What is the simplest way to do this, please?
  | 
  | Use md5sum

 yes (mostly) & no (just a little bit):  md5sum(1) can be
 used to tell you, with a high degree of certainty, which
 files are identical.  it will not, however, _prove_ any
 two files with the same MD5 checksum are identical.

 if you want to be absolutely (100%) certain two files are
 identical, transfer a copy of the file from one machine
 to the other, and cmp(1) them.  if `cmp' thinks they are
 identical, they are.
 ( for the paranoid, re-`md5sum'ing the transferred copy
  is a useful and quick check the file was not corrupted
  during the transfer.  reconfirming is important if you
  have a dodgy transfer medium/method. )

 checksum (MD5 or other) collisions are not impossible, but
 MD5 collisions are very very rare.  another tool, cksum(1),
 which computes the CRC, includes in its output the file's
 size.  that helps by making it very unlikely you have two
 different files with the same CRC.

 I have been known to first use md5sum(1) to find candidate
 identical files, then cksum(1) as a check, and then cmp(1)
 as a final check.  (b.t.w., the ancient sum(1) is rather
 useless and should be avoided.)

cheers!
	-blf-
-- 
Experienced (20+ yrs) kernel/software Eng: | Brian Foster   Montpellier,
 • Unix, embedded, &tc;  • Linux;  • doc;  | blf at utvinternet.ie   FRANCE
 • IDL, automated testing, process, &tc.   |  Stop E$$o (ExxonMobile)!
Résumé (CV) http://www.blf.utvinternet.ie  |     http://www.stopesso.com