Replied new files

Shell Programming and Scripting
1

We need to create a process that run continues and whenever a file is placed in a specific directory, the process detected this new file and replicated in another directory.
How can we implement a process that allows automatic replies to the new files?

Regards

2

take a look at the rsync utility.

3

Option 1: Use rsync from one directory to the other
Option 2: Use File::Monitor
Option 3: If you're on Linux, code something up in C using inotify
Option 4: (might miss files)

cd /incoming
while true
do
    filename=$( basename $( ls -1t ) )
    if [ ! -e /replicate/${filename} ]
    then
        cp /incoming/${filename} /replicate/.
    fi
    sleep 30
done
4

Tks for your awnser.

But it occurred one error with basename, it returns the fowlling error :

Usage: basename string [suffix]

Any ideia ?

5

Obviously 'ls -1t' in 'filename=$( basename $( ls -1t ) )' returns more than one file name, a thing which basename can't handle with. Try:

cd /incoming
while filename in "$(ls -1t)"
do
     if [ ! -e "/replicate/${filename}" ]
...

P.S. I didn't use basename at all, since /incoming may have subdirectories with files, which also should be taken into account while generating replicates

6

Wich values you will pass to filename ?

while filename in "$(ls -1t)"

7

I've made an error no double quotes are needed around $(ls -t)

To show, what values the variable filename gets, here is my example:
>cd sample_dir/

>ls -t
file3
file2
file1

> for filename in $(ls -t) ; do echo "filename is $filename" ; done
filename is file3
filename is file2
filename is file1

8

It work's,
one last question : How can i only list files from a directory (no need the files in the subdirectory), ie list files and exclude the directorys.

regards

9

in such cases I'm usually using 'find' command. Below is one possible solution. (Here we need basename in one place, since 'find' adds relative path in front of found files)

# your source/target directory may change
# so save it better in a variable
SRC_DIR="/incoming/"
TRG_DIR="/replicate/"

# looking for all files in source directory
# dereferencing soft links, not going deeper in
# subdirs - just as you wanted :-)
for filename in $(find -L $SRC_DIR -maxdepth 1 -type f)
do
    # looking if corresponding file does not exist in target directory
    if ! test -e "${TRG_DIR}$(basename $filename)"
    then
        # copy if it doesn't
        cp ${filename}  ${TRG_DIR}
    fi
    # don't sleep on your workplace :-)
    #sleep 30
done

I've commented 'sleep' because I think that it would be better to add the execution of your script to crontab.