UNIX for Dummies Questions & Answers

Hi,

setsize is the total number of files divided by the number of parallel processes. I added 1 so that you will have no more than the number op parallel process. In this case the workload is split in 3 and 1, but for larger numbers it works out more evenly.

wc -w tests the number of entries although wc -l is perhaps more appropriate in this case
xargs cuts up the set of files in more or less equal parts
each cp process copies the files in these equal parts to the target directory.

The & puts the copy processes in the background and the wait command waits for them to finish. So the are not multiple threads but multiple processes.

The script breaks with file names that contain spaces. I'll fix that when I have more time.

S.

Code:

wc -w

is used to get total number of files.

The tricky part I did not understand and I am waiting for Scrutinizer is the code

Code:

| xargs -n $setsize |

How does this divide the file set and spawn appropriate parallel thread ?

I am learning how to use this forum. I can now use code tags

---------- Post updated at 02:46 PM ---------- Previous update was at 02:40 PM ----------

Scrutinzer,
Thanx for your explanation. I just saw your response it flipped over to page two and I was waiting like a fool on page one.

Question can you please explain :-
So the are not multiple threads but multiple processes.

So what is the difference between a thread and a process ? Are they not the same ?

Multiple threads means multiple execution contexts inside the same process. This is less weight on the operating system since it doesn't have to switch process contexts as often, but is more difficult on the programmer since they have to worry about race conditions etc. You can't do threads from the shell.

Multiple processes is multiple processes.

They can both run several things in parallel.

I would use a tool like rsync or unison that are optimized for this kind of use and making backups.

Try this:

rsync -avz /data4/dbx /backup/dbx