Experimenting with VMs is the way forward.

Basic networking knowledge is vital. And being able to configure your own firewall(s) safely is an important skill. Check out something like Foomuuri, or Firewald. Shorewall is brilliant for documentation and description of issues (with diagrams!) but it does not use the newer Linux kernel nftables and is no longer actively developed.

Go for it with Nextcloud.

I would also recommend at least having a shot at setting up an email server, although I would recommend pushing through to a fully working system. It is possible, and is very satisfying to have in place. The process of setting one up touches so many different parts of internet function and culture that it is worth it even if you don’t end up with a production system. The Workaround.org ISPMail stuff is a good starting point, and includes some helpful background information at every stage, enough so you can begin to understand what’s going on in the background and why certain choices are being made - even if you disagree with the decisions.

Python is great for server admin, although most server config and startup shutdown snippets are written in BASH. You will no doubt have already begun picking that up as you interact with your VMs.

I came here to upvote the post that mentions haproxy, but I can’t see it, so I’m resorting to writing one!

Haproxy is super fast, highly configurable, and if you don’t have the config nailed down just right won’t start so you know you’ve messed something up right away :-)

It will handle encryption too, so you don’t need to bother changing the config on your internal server, just tweak your firewall rules to let whatever box you have haproxy running on (you have a DMZ, right?) see the server, and you are good to go.

Google and stackexchange are your friends for config snippets. And I find the actual documentation is good too.

Configure it with certificates from let’s encrypt and you are off to the races.

Working here (UK).

And will be cancelled in 18 months with 2 weeks notice.

This approach sounds good.

I think the correct approach is both, if you have the option.

Most devices accept two name servers. Redundancy is always good, especially for DNS.

If you weren’t at a university it was generally a challenge to get hold of disks. Downloading at home took forever on a 28.8 or even 56k modem (ie. 56 kilobits per second).

Slackware and Redhat disk sets were the thing, in my experience. But generally that only gave you the compiled code, not the source (although there was an another set of disks with the source packages).

If you wanted to recompile stuff you had to download the right set of packages, and be prepared to handle version conflicts on your own (with mailing list and usenet support).

Recompiling the kernel with specific patches for graphics cards, sound cards, modems and other devices (I remember scanners in particular), or specific combinations of hardware was relatively common. “Use the source, Luke!” was a common admonition. Often times specific FAQ pages or howtos would be made available for software packages, including games.

XFree86 was very powerful on hardware it supported, but was very finnicky. See the other posts about the level of detail that had to be supplied to get combinations of graphics cards and monitors working without the appearance of magic smoke.

Running Linux was mostly a enthusiast/hobbyist/geek thing, for those who wanted to see what was possible, and those who wanted to tinker with something approaching Unix, and those who wanted to stretch the limits of what their hardware could do.

Many of those enthusiasts and hobbyists and geeks discovered that Linux could do far more than anyone previously had been prepared to admit or realise. They, and others like them, took it with them into progressively more significant, and valuable projects, and it began to take over the world.