Recovery Reveals Another Bad Disk
The reason the previous reboot lost all of the containers seems to be because it lost one of its disks.
There was a disk that reported it would need to be checked for errors on reboot. It's been there for a while, that warning, and it seems the check on reboot either didn't find anything or couldn't fix it.
At the last note, after I restarted following a normal system update, I had to re-add all of the Docker containers. After I finished that and took a little breather, I tried to look into why. I quickly discovered the /var folder was nearly empty, as the partition it should be on didn't mount on boot.
Since it wasn't mounted, I was able to runĀ fsck on the command line. It found errors, and tried to fix them, but ended with a "cannot write," after which any disk peeking tool like parted or lsblk failed to see the drive.
I crossed my fingers and rebooted, thinking maybe a run through the BIOS might tickle it, and whatever state was maintained by the running OS would go through its discovery and hopefully at least recognize the disk again. Instead, it's gone completely, and even lshw doesn't see it.
The reboot did see all of the containers again, because they're in the /var folder on the root partition. I noticed I ran some of them without specifying they should restart, so I quickly restarted them; I'll add that to my notes and correct them tomorrow.