This script tarball was the direct result of budget cuts to a publishing department.

Using old 486's ( in this case SX25's with 8 meg of ram) tied together on a standalone network, these shell scripts assemble to effectively become a diskcopy machine.
The idea is to duplicate a disk image and store it locally. This image can then be dumped onto fresh disks when needed.

Requirements

One monitor, one keyboard, network gear and as many machines (doorstops) as deemed necessary

The "master" computer may need to be a (minimum) 486 50MHz 16 meg ram machine for sanity's sake.

The linux version I used was RedHat 6.2, squeezed down to about 60 meg to fit on the small disks we had available. Anything larger in the processor, memory or hard disk department will make life easier, but they are not essential. ANY linux should work.

The number of machines is entirely at your discretion, it could expand to ridiculous limits. In our setup, 10 machines gave 20 available floppy drives, which was more than enough to keep one person busy loading, unloading and labeling the results.

There is a distinct lack of documentation with the package, however if the flopycopy.tar.gz is unpacked on each machine, rsh enabled and the network configured then it should be reasonably straight forward to get working.

At the start of a copy run, each machine is polled for availability and a warning given. The operator then needs to take the warning message on board.
If there are no disks inserted in a remote (client) machine, it will be ignored for that run.
Disks must be present in the master (server) machine for copying to progress

Ctrl-C is your friend.

/etc/hosts needs to know about all your client machines

These machines are a security risk if enabled on anything other than their own STANDALONE network. On their own network they are fine.

about.sh gives an overview of the menu options

install_info contains some notes regarding the initial machine configuration, they were jottings for myself and the particular setup I was doing so not all the notes are relevant

This was written as a quick fix, and had no budget for development. It's probably got lots of bugs but in the actual duplication area the copies made are sound. There was a marked reduction in the return rate of disks after the switch was made to this process, as every disk is checked against the original image before removal.
It could do with some usability enhancements and a code clean up no doubt.

rsh appeared to play up on one batch of machines - cause unknown. Lesson learnt:- don't fiddle with a working setup!

The remote image directories may fill up depending on usage, provision needs to be made to periodically clean them.

The department has been and gone, the machines are landfill somewhere.
If you can use the code and need some help then I can offer very little, except to improve the documentation somewhat or fill in a glaring omission in this quick spiel.

If the department had survived then I would've devoted more time to converting it all to a floppy based distribution using rsync, even perhaps using etherboot for the client machines, but that's all ancient history now.

If it's useful, it yours. The GPL notwithstanding.

One interesting outcome of this effort, besides the huge cost saving, was that inventory levels were managed a lot better. When stock levels can so easily be topped up there's no need to run off excess copies!