Mismanagement on my side. I was in the process of hardening my node (I should have done _all_ of this at the beginning) and (don't tell anyone) locked myself out. Luckily, I had a snapshot of the VM just a few minutes back and decided to shut down and roll back. At this point, I still had the real-time backup file (lightningd.sqlite3). I then started lightning, assuming it would either go with the most recent version or refuse to start because the DB and backup DB are not identical, at which point I would have duplicated the backup DB. Instead, lightning started, complained bitterly ( **BROKEN** lightningd: backtrace) but continued operation. I thought it would heal itself. I assume I should have duplicated the backup file _before_ starting lightningd. I am not worried about the funds that will come back (minus fees).