I was having trouble when trying to run this with it halting in each Pool worker waiting for domain input.

To get around this I've added a main() function that runs only in the entry process.

I've moved all the code not already in functions into this function. This took some variables out of global scope so I've also refactored the functions to be passed the variables they were using from global scope where appropriate.

There are a few changes here that don't change the functionality (except in one case).

Using context managers

I've changed the code to use context managers for using the multiprocessing pool and for writing out the files to disk, this shouldn't affect how the program works but removes the need to call the tidying up functions.

Ran black

I've run both python files through black, just to make it easier to keep a consistent code style.

Added a WaybackRecord data class and a parsing function

The getUrls function was doing a bit of record parsing and I wanted to extract this for a few reasons. I think it improves readability and may improve the ability to add some filtering to what is downloaded before it is downloaded in the future. A next step there might be separating the creation of wayback URLs from those records.

Extracted the filename creation function

I've added a function called safe_filename to create the filenames that content is downloaded to, extracting this means in the future it could be unit tested separately in the future and hopefully makes download a little easier to read.

Refactored the retry logic in the download function.

There is a check that the output filename isn't over 255 characters, I've move this outside of the retry logic as it the filename doesn't change between tries.

I've changed the type of loop to a for loop so the number of retries can be limited just in case it gets stuck in a constant retry situation.

I also move the file writing code outside of the retry loop as I don't think that needs to be in there.