I am trying to write unix script which will help to move files from dir A to dir B.
dir A will get files from ftp connection.

ftp >> dir A >> dir B

I have no issue to second part i.e. move files from dir A >> dir B

My doubt is if I can use lsof command to list out all files which are completely ftp'd to dir A?

and do lsof command can check on completely downloaded files and ommit partial downloaded at dir A?

linux commands come packaged with man pages that are manuals for each command's use

they also describe what the command does

just type man in front of the command name..


so standalone lsof will show all open files on the system..
you could grep out a specific directory

basically while the file is downloading you will see a lot of results in there from all of the different services using the file causing it to be in an open state

once the file is completely downloaded, or the download fails and ceases,
the output of lsof on that directory will be empty

theoretically you could write a bash script that watches a dir and determines if the output of lsof is populated or empty and assign actions for each outcome:
populated : check again in 17 seconds
empty : alert the file has completed downloading

this is useless in determining if the file completely downloaded,
for that you would have to write another part of the script that checks the size of the file on your system to see if it matches the size of the file on the ftp server

also it would be unnecessarily slow,
it would be better to write a script that performs the get and then when the get finishes alerts you with each file

1 members found this post helpful.

Check out the "wget" utility which handles FTP. IT has some methods of handling incomplete transfer, and suitable exit codes to indicate a failure.

Well, no, that's not what lsof does:
More typically you would simply use ls to see the list of files. Command line options for ls that might be useful would be
See the manual pages for both utilities for more.

I do wonder why you are using ftp to copy or move files from one directory to another; the cp utility (or the mv utility) can do the same thing, possibly more quickly.

You would use cp -pr dirA dirB to recursively copy the content of dirA to dirB. You can also simply rename (with the mv utility) dirA to dirB and not copy anything if both directories are in the same disk partition (if they are in separate disk partitions mv will do the recursive copy for you); across a network interface it's a little more complicated, however.

Hope this helps some.

No actually I am getting files from other server through ftp connection @ dirA
So far now I have write script like this:

CHECK=`ls -al dirA |grep flagfile |wc -l`
cd dirA
touch dirA/flagfile
echo `ls -ltr`> HOMEDIR/result1.txt
sleep 5
echo `ls -ltr`> HOMEDIR/result2.txt
diff HOMEDIR/result1.txt HOMEDIR/result2.txt > HOMEDIR/compare.txt
COMPARE=`cat HOMEDIR/compare.txt|wc -l`
if this is compare is "0" it means my file is ready to move at dirB

but what if I don't want to spend time on "sleep 5". In fact I can afford to lose this time..
as there would be thousands of files floating.

Is there any alternate way I can do that speedily.

Why are you doing that?

Like I indicated earlier - use wget. When it finishes with a successful transfer, move the file.

Not sure why you suggesting me to use WGET whereas I am not downloading any files WEB.
file processing happening like this.

Other server >> ftp'd files >> dirA >> dirB

dirA and dirB are same unix server.

since we have to use ftp I am wondering how can save my time to check if files are completely doenloaded at dir A from other server. (through ftp)

Sigh. Wget is just a file retrieval utility that supports FTP and other protocols. The point is that you can use it to obtain the file you need and check that the transfer was successful, presumably by checking the return code. If you're using Bash, this will be stored in the $? variable.