Need help for writing shell script

[chris_87]

Hey guys, I am a new baby here, I need some help in writing a shell script, it basically need to compare the two file contents in xyz time intervals, if the size of source folder is larger, then it will automatically rotate the file to the destination folder, and rename it to folder.1 folder.2, folder.3..etc without overwritting the folder in the destination. Can anyone help me?

thanks

[themachine]

Perhaps you could be a bit more descriptive... I got a little confused there. Are we talking about *files* or *directories*?

Maybe if you explained to purpose a little I could give a better example.

[themachine]

Ok, so I played around a little... I don't know how well this will help, but here you go:

#!/bin/bash

# simple script... bla blah

BASE_DIR=/home/johnny/bin/test
DST_DIR=$BASE_DIR/dir
SRC_DIR=$BASE_DIR/src_dir
ROTATED_SUFFIX=`date +"%m%d%y-%H%M%S"`


# Move to the base dir
cd $BASE_DIR


# You probably want to test for a newer directory rather than a bigger one
# -nt = Newer Than
# You can modify the script to use this condition rather than the 
# Directory Size

if [ $SRC_DIR -nt $DST_DIR ]; then
        echo "$SRC_DIR is newer than $DST_DIR... It should be rotated!"
        echo
else
        echo "$SRC_DIR is not newer than $DST_DIR... Don't rotate!"
        echo
fi


# "-e" Tests that $DST_DIR Exists and then gets the 
# byte count for the directory
# du -cbs = "Disk Usage -c[total] -b[bytes] -s[supress output]
# awk gives you just the bytes count, nothing else

if [ -e $DST_DIR ]; then
        DST_BYTES=`du -cbs $DST_DIR | awk '/total/ {print $1}'`
        echo "The destination directory has $DST_BYTES bytes"
else
        echo "WTF => $DST_DIR doesn't exist!"
        exit 1
fi


# Test for $SRC_DIR and then get the byte count for the directory

if [ -e $SRC_DIR ]; then
        SRC_BYTES=`du -cbs $SRC_DIR | awk '/total/ {print $1}'`
        echo "The source directory has $SRC_BYTES bytes"
else
        echo "WTF => $SRC_DIR doesn't exist!"
        exit 1
fi


# Test to see if the Source Dir is "Greater Than" the Destination Dir
# -gt = Greater Than
# If the SRC_DIR is Greater Than the DST_DIR then it moves
# The DST_DIR to DST_DIR.[datestamp]-[timestamp]
# and then copies the SRC_DIR to DST_DIR

if [ $SRC_BYTES -gt $DST_BYTES ]; then
        echo "The source directory is bigger!"
        echo
        echo "Rotating ${DST_DIR} to ${DST_DIR}.${ROTATED_SUFFIX}!"
        mv ${DST_DIR} ${DST_DIR}.${ROTATED_SUFFIX}
        mv $SRC_DIR $DST_DIR
else
        echo "$SRC_DIR is not greater than $DST_DIR... Not rotating!"
fi

P.S. Bash isn't my strong suit, but you know... a couple minutes and a came up with that... hope it helps.

[themachine]

If you want it to run say, every hour you could create a cron tab by doing:

Edit your crontab:
# crontab -e


Type this in the file:
1 * * * * /root/scripts/freshen_dir.sh 2>&1 >> /var/log/freshen_dir.log

You would of course need to name this script "freshen_dir.sh", and mv it to /root/scripts.... change that to whatever you want.


The "2>&1 >> /var/log/freshen_dir.log" will send all output (errors and std_out) to a log file for review.

Cheers!

[chris_87]

Thank you for your kindly reply, let me try to explain my idea clearer.

Basically I am going to setup a ftp server, setup one account for incoming files with no download read and delete rights, and setup another for my personal account.

Once my friend upload files to my 'incoming' account, I want to write a script that would able to compare the 'incoming' folder in the 'incoming account' (ie /home/annoymous/uploader1/incoming) to the 'incoming' foder in my personal account (ie /home/ftp/my_own_personal_account/incoming), if the size of source folder is larger than the destination, it would copy the files to the destination folder, and do nothing if it is smaller.

I want to do this because I've got some other ftp user, I know I can do it easly from the shell, but the other ftp users have no access to the shell, so I want it to be done automatically for serveral intervals (say 2 times a day).

Besides, I also need help to write a script that help me the clear all files in a folder (ie /home/annoymous/uploader1/incoming OR /home/annymous/uploader*/incoming in batches), but in this case I will only need to to run manually.

Many thanks again of reading my 'long' question and I really appreciate your help.

[themachine]

In that case... with little modification that script i posted should be able to work for you... I would recommend using the "timestamp" check rather than checking to see if it is larger... mainly because the newer files could be smaller in size therefore it wouldn't work.

As far as deleting everything in a directory.... you could just throw that into a cronjob... i.e. -

1 * * * * /bin/rm -rf /path/to/your/directory/* 2>&1 > /dev/null

Put that in the appropriate users crontab (as root rype 'crontab -e -u <username>' and that is all you would need. That line right there will delete everything in '/path/to/your/directory/' but not "directory" itself.

Hope that helps.

[ckp212002]

themachine,

i am a newbie too, can u pls explain the following line:

DST_BYTES=`du -cbs $DST_DIR | awk '/total/ {print $1}'`


thanks

[drks]

Originally Posted by ckp212002

themachine,

i am a newbie too, can u pls explain the following line:

DST_BYTES=`du -cbs $DST_DIR | awk '/total/ {print $1}'`


thanks

Certainly. Let me break this down:

DST_BYTES -> The variable we are setting. In Bash scripts you do not set a variable with the leading '$', however you do reference it as '$DST_BYTES' after you set.

`` -> The Backtick '`' (the one on the same key as tilde '~') signifies the execution of code. Bash will spawn a sub-shell, and anything within the back ticks will be executed. The output of the command is sent back, and becomes the value of our variable 'DST_BYTES'.

du -cbs $DST_DIR -> The 'du' bash command is used to show 'disk usage'. Therefore we are checking the disk usage of the directory $DST_DIR.

| -> The Bash pipe operator '|' (on the same key as the backspace) is used to pipe the output of the first command, and send it as arguments of the next command (awk).

awk '/total/ {print $1}' -> Awk is a beast to grand to explain in a simple one liner. However this command is doing a regular expression match for the line that has 'total' in it. Once it has that line, it prints the first ($1) word in that line.