I am using pax to backup my files and need to exclude a couple of large files. The file names never change.

From reading the man pages i should use the -c operator but it doesn't give any examples. this is my best guess but it says i am using a invaild combination of w c and f

pax -w -f test /root/code_samples -c /root/code_samples/phpMyAdmin-2.6.2.zip

Does anyone know how to exclude a file??

thanks

t

I think pax could turn out to be a nice little gem and may have to check it out some more.
The man page can be a bit confusing for sure but here is an excellent article. The -c option is on page three and is used like this.... -c './Linux Stuff' where that is the directory which I don't want restored.
http://www.onlamp.com/pub/a/bsd/2002...cs.html?page=1

Here is an example I used.
Don't forget to use single quotes on the file or directory to exclude from the restore.

so the -c flag is only good for excluding a folder from being restored?
Not excluding it from being in the backup in the first place?

That's what I gather so far. I'll have to do some more checking to see what other options there are but I don't think -c is used during save operations.

what i want to do it to backup a users site into a single file and then offer it back to them so that they can download it and save it. (Basically so they are ultimately responsible for the data and not me)

My users like many others are a bit thick and we have advertised as a 'no extra software required' websites. So i would like to avoid using a FTP client. It would be great if the user just clicks a link and off they go.

Therefore i thought if i can backup /home/website1 and then copy that file to /home/website1/backup the users can download from /home/website1/backup directly (obviously with sufficent authentication)

This leads to the problem that the next time the backup is run it will backup the backup. Because the user is downloading the backup file it is vital that it is as small as possible.

Can specific file be removed from the backup file after creation??

You are about to find out just how "thick" these people can be.

I don't know what all options you can use.

the man reason i am doing this is because one of my users deleted all the products out of the database then phoned me asking what the undo command was. He wasn't happy when i said there wasn't one. I did have a backup but it was before he added 400 products, again he wasn't happy.

I am open to any and all suggestions

thanks for your help

Okay this is involved but not too complicated. The find command is going to find every single file in their directory. It will pass that to the grep command, which will remove the lines which match the pattern of the file you don't want (use the whole name and path -- look at the find output to determine what this is -- to make sure you get only it). The grep command pipes all those filenames to pax... which writes an archive. The -d flag is what is important. It tells pax to match directories but not all the files beneath it. Basically, it says "only archive what I specifically name". In this case I use ustar as my format but you can use whichever makes sense for your users and is supported. I also pass it through bzip2 to compress it as well. This may or may not work for you (if you don't think your users will have bunzip2 or access to a program that will handle it).

Okay, here it is in more of a script format.
Note, you could pass each user's directory as an argument and just assign it to DIRECTORY, or you could modify this script to step through each one... and do this for each. But it gives you the basic idea.