Every day i ftp tar.gz a file from the production server to a back up machine.. This task creates way to much traffic on the network at the end of the day and puts and undo load on the production machine during operation hours. i would like to create a script that would automatically fire off the ftp at like 2:00 am so that the transfer occurs when there is little or no load on the network. if anyone has suggestions on how to accomplish this, PLEASE let me know...

thanks..

--e0

loworderbit@aol.com

You can set up a ".netrc" file for the
user doing the transfer then set up a cron
job for that userid to run a script to actually
do the FTP transfer.
Check the man pages for ftp(1) and netrc(5)

I need to do a lot of automated ftp jobs myself. I have tried several versions of this and have finally settled on a style of script that I like. I found that using a .netrc file to automate the logging-in process kept painting me into a corner because different scripts needed to sign in as different users. So I avoid .netrc and force the script to sign in. I don't like to allocate pty's unless I really am forced into it, so I also avoid pty based tools like expect. I really like ksh so that was my tool of choice. And I like the co-process concept because it makes feeding commands into the ftp process so easy with "print -p". The only problem is that the co-process manipulates standard-out so as to make it available to "read -p". And it's too hard to know how many "read -p" I will need. So I send the output to a different file descriptor. Putting it all together:

That script will tranfer the file and the output of the script will be the output from the ftp job itself. Put the script into cron and save the output so you can look at it the next morning.

I'm using your code but i have a problem,
i need the code to be flexible in such a way that it changes the filename it gets everyday. And it should be done automatically.

For example,
1st day: get apr20.log
2nd day: get apr21.log

hmm... write a script on the server that is launched daily using the cron (man crontab)... have that script ftp to your workstation or wherever and grab the file. The automated FTP will probably require a .netrc file in your user directory or wherever you ftp to to get the file.

However, if you are having to tar the file manually every day, write a script on the workstation side to do that for you a few minutes before the server script executes... once again, utilizing the crontab.

Good luck and let me know how it goes.

Strange... those other replies both popped up while I was writing... ummm... take their advice.

Perderabo... can you explain what that does exactly?

That is the file descriptor manipulation. Recall that fd 0 is standard-in, fd 1 is standard-out, and fd 2 is standard-error. The line "exec 4&>1" opens fd 4 and assigns it to whatever fd 1 was assigned to. As you will see, I am sorta "saving a copy of fd 1 in fd 4".

The line "ftp -nv >&4 2>&1 |&" is a little harder.

The "|&" turns the process into a co-process that allows subsequent "print -p" statements to send lines to the co-process' standard-in and "read -p" to read from its standard-out. So ksh forks a copy of itself and fiddles with the fd's 0 and 1 until this it set-up. But it leaves the rest of the fd's alone.

Then it encounters ">&4" which causes it to set the ftp process' standard out to whatever 4 is. Well since 4 is a copy of 1 before the co-process, we are back to writing to the original shell's standard out. Lastly, the 2>&4 does the same thing for standard error. I could've used "2>&1" at this point for the same effect.

This is hard to explain, but I hope this helps.