There was an outage lasting a couple of days. Service should now be restored. The root cause is misadventure. However, it is more accurately described in the literature as an own goal, arising from the reckless deletion of the files that make-up this blog and an inadequate approach to backups.
Site Admin
This site has migrated away from Cloudflare. While Cloudflare was overkill for this site, the push to incorporate and support more AI along with the recent announcement to sack over 1,000 workers to ‘invest’ in AI was the driving factor.
Domain hosting is now with Porkbun. The service is intuitive, accessible and works. All that is needed from a domain register. I’ve used Porkbun for a variety of domains for over a decade.
The DNS and CDN is now provided by Bunny. While a CDN is probably overkill for this site, I like the fact that it reduces load on the server.
This migration went well.
Misadventures in optimisation
So far, so good. The next phase was to optimise the site post migration. Involving poking around the server, running update scripts, and cleaning up errant files. I use the app Secure ShellFish as it integrates into the operating system. I can edit and upload files directly from the Finder.
However, the migration resulted in a sync bug and duplicate folders. I like to keep the directories tidy. So I deleted some of the duplicates. Deletion on the remote sever is the real deal. There is no undo. The files were the site.
Reconstruction
Where is the backup?
Having a backup is essential to disaster recovery. Having a backup that is an exact copy of your data at the time of incident is invaluable. Reinstall, upload, build – you’re good to go.
Various sources
My development approach not systematic. There was no ideal backup. This involved recreating the site from various sources including downloading content from cached browser tabs and the dev files.
There was enough raw material to spend the past couple of days rebuilding the site. Getting the feed to work has been a challenge.1
Next
It is great to be back up and running. The rebuild was the refresher into how this site is assembled and the various components across the stack.
One takeaway is the need to develop a more robust workflow. This may include having to learn Git to allow for a CI/CD pipeline. I want to stay away from GitHub because their wanton approach to copyright and protection of customer data in their repositories. A limiting factor is that this locks development to the laptop.2
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.