Anyone fancy giving me some scripting help?

I have a Linux machine where I have a lot of folders within a folder and I need to make a tar.gz of each of the subfolders and then delete the original subfolder Something like


After the script I end up with


Eventually, I'll probably tar.gz the resulting files as well, ending up with only one file.
I tried yesterday to select all the folders, right-clicking and selecting Create archive. After it had run for about 16 hours, it had done about 2/3 of the selected folders. Seems it will tar.gz one folder, add it to the finished result, tar.gz the next folder, untar the result before adding the second folder. As the tar-file grows, it will take longer and longer time to add one folder. When I stopped it, the resulting file was 12Gb and 178k files...

I think this script will be quite simple, butI'm not very adept at scripting...

How about this?



(commented out the rm -r - only uncomment when you are sure everything works the way you want!)

Thanks, Julf.

That worked. I removed the verbose output from tar and added a line above it instead (echo "Archiving $subdir"). It got a bit old watching 170k+ lines scroll across the screen...

Ah, yes, should have removed that after testing...

Note that the script is not very robust - it is a typical 2-minute off-the-cuff thing...

Still, it worked for what I needed.

It looks more robust than what I'd have suggested:

gzipping a tar file that contains only other gzipped files is going to gain you very little beyond using up some CPU time.

You'll probably actually lose a miniscule amount of hard drive space due to the overhead of the zipping program, but you will have a cleaner-looking directory tree.

To some of us OCD types, that is worthwhile.

tanstaafl.

You won't have a cleaner tree by doing the extra gzip. You can just have the top level file be a tar file, rather than a gziped tar file.

(you don't really have the equivalent of a tar file in the Windows world, a tar file is kind of like a zip file but with no compression)

I may not gain anything by gzipping the resulting files, but it's easier to just select all the files and select Create archive than to do the commands in the shell. I don't really have much problems using the shell, but my collegues on the opposite crew aren't very Linux-savvy.

Can you not just do a tarred but not zipped archive via the same route ?

If you mean using the file browser? Don't know. Maybe by changing the extension when inputting the file name, but it works this way as well. That's good enough for me...