Thanks. I did reply to your email.
You might have misunderstood what I said. I haven't spent 4 months fixing the server issue. I spent 4 months doing other things that were more important. I started setting up the new box on Saturday afternoon and had a test forum up and running by yesterday morning.

I wasn't aware you offered dedicated self-managed root servers running within containers under Plesk? (required as there are multiple webspaces, not the same as multiple domains, and almost 2 million inodes)
The forum software is all written in perl and the pages are dynamically created by building arrays. There is no 'database'.
Seeing as some of the array functions are now deprecated in the latest perl, I had to effectively construct my own modified version of Ubuntu 20 to process these functions outside of Apache. Maybe took 6 hours.

The only reason I'm suddenly doing this now is I got a reminder that CentOS 6 support ends on 30 November, and CentOS 8 does not support certain perl modules any more.

The https issue is my choice. Of course I could flip it over at any time but there are 1.2 million archived posts indexed on google and I can tell you the indexing will die.
Similarly, each archived post contains the full URL of any uploaded image within it, which is as I explained above dynamically created. Therefore setting for example all http to port to https won't work.
I know all that from the https testing I did yesterday before reverting back to http.