So here's what happened. Today I was bored and went to check mix-space to see if there were any backend updates. I noticed there was a v10 update, so I went straight to the console and updated it.

After the update, I felt it wasn't mature enough and wanted to roll back to v9.7. But after rolling back, I couldn't log in no matter what. I suspect the database version had been updated, making it impossible to downgrade.

Then, after a bunch of clueless tinkering, I successfully trashed the backend database (last time I trashed someone else's unified authentication database, and this time I trashed my own main site's database — what a talent!). After that, the backend was inaccessible, and the frontend couldn't connect to the backend, throwing errors everywhere.

That's when I suddenly remembered that I had set up 1Panel to do a daily scheduled backup of the entire main site Docker Compose directory (including both frontend and backend). I rushed to my NAS, transferred the backup archive to the server, rebuilt the image, and brought it back from the dead. Nearly gave me a heart attack.

In the end, I decided to go into 1Panel and change the backup execution interval from "daily backup" to "every eight hours," and increased the number of retained copies from 3 to 30. I was literally scared to tears.