backing up text files distributed throughout the drive
 

Hi,

I am attempting to back up a bunch of .m files that are located all over the 2 hard drives (these .m files are text files run by matlab). I have been doing a bit of research and its seems like cpio might be a good option.

Is this the command that I would use?

find -name "*.m" -print | cpio -ov > mfiles2007.cpio

Is there a better way to accomplish what I am setting out to do?

Thanks for any help you can offer me.

Paul

My advise: Test the command out on a test collection first, just to be safe. Other then that i cant say, as cpio has no man page because GNU doesn't like man pages, so i have no clue if that command will work.

It will work if you can guarantee that no directories or file names have spaces in them, unless CPIO handles spaces in stdin streams....

I'd use:
locate -r '\.mp3$' > filelist.txt
for line in $(cat filelist.txt); do
cp -v "$line" /some/destination/
done 2>&1 | tee copy_report.txt

I'm sorry, I don't really understand this:

locate -r '\.mp3$' > filelist.txt
for line in $(cat filelist.txt); do
cp -v "$line" /some/destination/
done 2>&1 | tee copy_report.txt

Is this 2 commands separated by the semi-colon? Also, I only want .m files not .mp3's but I can figure out the substitution to make there myself :)

Once again, I apologize for my ignorance,

Paul

"find /data -name "*.ogg" -print | cpio -ov > test.cpio"

I just ran this against the ogg files in my data disk. They have embedded spaces, and it seemed to work just fine. I did a "cpio -tv < test.cpio", and it listed the files I expected, having substituted "\ " for the embedded spaces. Sometimes you just have to do it and see if it works. :)

Added:
Almost forgot: the man page sucks, but "cpio --help" or "cpio --usage" gives a lot of help.

The 'find <options> | cpio <options>' is the most powerful, but also two of the oldest and most archaic commands in /usr/bin/*.

Let me break down my method:

locate -r '\.mp3$' > filelist.txt
# There is a database updated nightly of all filenames.
# The above makes a list of all files ending in '.mp3'
# The advantage is that you can just edit the list. b)

for line in $(cat filelist.txt); do
# This assigns each line of filelist.txt to $line.
# The semicolon is just a stylistic shorthand.

cp -v "$line" /some/destination/
# The meat of the command.
# I actually run 'cp -vn' first which
# pretends to copy, but doesn't actually do it.

done 2>&1 | tee copy_report.txt
# Some UNIX magic here. This ends the looping
# over filenames.txt and forks the output of
# 'cp -v' to both the screen and a logfile.