Shell Programming and Scripting

Hi all

I have a unix based firewall, which creates a daily backup file on the device.

I need a script to scp this file over to a remote server.

I can get this working daily using a basic script and a cron job.

However, I only want it to send the latest config back up file and currently it is sending across everything is the source directory.

How can I amend the script to just send the latest?

Thanks




James

Welcome jimmyzoom ,

I have a few to questions pose in response first:-Most importantly, What have you tried so far?

There are probably many ways to achieve most tasks, so giving us an idea of your style and thoughts will help us guide you to an answer most suitable to you so you can adjust it to suit your needs in future.


Without seeing your script, it is difficult to suggest how to change it. Can you post it? Please wrap it in CODE tags so it displays properly.


Kind regards,
Robin

If you use a standard tool like rsync, you can have it only backup files in the src directory that are new, meaning that were not there the last time the backup was run. If your "latest config back up file" has a different name than the files that have already been copied, rsync (or many other backup programs) will automatically do what it sounds like you want. If the "latest config back up file" has the same name as a file that already exists in the mirror location, rsync will rename the older file (the one in the mirror location) with an extension you provide. If you want to have multiple incremental backups of a file with the same name, that can probably be done as well but I prefer to rename older versions of a file for clarity and book keeping purposes.

As rbatte1 suggests, more information would be helpful. At any rate, a backup solution should be designed with your recovery process in mind. What files you store, and where, and what they are called, etc, depends somewhat on the process you will need to involve should you ever have to use your backup files.

You will want a script to run that end of the business as well, and definitely practice using it before you need to use it.

LMHmedchem

One more point, in addition to what has been said:

You need to define, what you mean by "latest". The one having the most recent modification time, or the one having a "most recent" timestamp embedded in the file name?

I found that in practical applications, the latter is used more often, because it is too easy changing the modification time of an old file to a new value by accidentally touching it. Writing a backup script of course depends on how you define the "latest" file.

Personally, i think that using dates (or other form of timestamps) in filenames is not necessarily a good idea: the advantage of knowing when it was created/last modified without relying on inode information is somewhat alleviated by the unability to restrict the creation of files to a certain number of generations.

The usual case is like this: create a daily backup and keep a certain number of generations (say: 5). One option is to create a script creating a timestamp-named output file and then another script to delete the 6th-oldest such file (if it exists). Run both scripts daily. Possible problem: the second script is oftenly forgotten and the backups fill up your diskspace.

Another option is to create a file with a fixed name and apply logic to rename this file until a name for deletion is reached after a certain number of generations:

Code:

#! /bin/ksh

if [ -e /some/where/4th_gen ] ; then
     mv /some/where/4th_gen /some/where/5th_gen
fi
if [ -e /some/where/3rd_gen ] ; then
     mv /some/where/3rd_gen /some/where/4th_gen
fi
if [ -e /some/where/2nd_gen ] ; then
     mv /some/where/2nd_gen /some/where/3rd_gen
fi
if [ -e /some/where/1st_gen ] ; then
     mv /some/where/1st_gen /some/where/2nd_gen
fi

my_backup_process > /some/where/1st_gen

exit 0

Advantage: you only need one script and you never have to bother with a process filling up your diskspace with numbers of old backups.

I hope this helps.

bakunin