…trying this again.

So, this blog is self-hosted, with a static site generator called Jekyll (and good old fashioned Apache as the web server). Originally, I had this hosted on virtaul system on Oracle Cloud. Things were working well, until, after a series of upgrades, frustated commands, and a reboot, I killed the server.

I couldn’t get it back up.

In a fit of frustration, I destroyed the server and spun up a new one. That’s what cloud servers are for, right?

That didn’t work.

A primordial scream later, I decided to set everything back up on a cloud server I had doing nothing hosted by SDF, my favorite retrocomputing site. My daughter invited me to go third-pace at Birdie’s Coffee Company. After an hour and a half of typing in a variety of commands, google searches, opening firewall ports, and mild frustation, I have most of it up.

There were four core sites hosted on the original VM. My process for three of those sites was to draft it locally on my MacBook, and save it to some cloud storage that was also mounted on the VM. Then, I’d copy it over. Those sites were restored quickly.

I did not appear to do that for one site. Unfortunately, it’s the site I tend to point most external folks to. I was able to go to the Wayback Machine to restore a copy from February. I’m going to have to clean up the links, but I think I’ll be able to restore it. It will just take a lot of search-and-replace. UPDATE A FEW HOURS LATER: I’ve restored it at least to what the Wayback Machine had.

There are a few scripts and other minor configuration and creature comforts I’ll have to rebuild. That shouldn’t take terribly long.

What did I learn?

  • Take your time when upgrading.
  • Take your time when troubleshooting a VM.
  • My “draft local and save to the cloud” process has value.
  • Back up your VMs.