An unclean shutdown can leave Home Assistant or one of its apps unable to start because the record describing its container is damaged. The Supervisor recognizes a specific form of this failure and removes the broken record so it can recreate the container, instead of failing on every startup attempt.
The problem followed a change in Docker, the software that runs those containers. It kept damaged records visible while rejecting normal operations on them, preventing the Supervisor from reaching its cleanup step. That could also block backups of an affected app or disrupt app authentication when Home Assistant Core was affected.
Stefan Agner's correction restores that recovery path and records which container had the problem. It addresses this particular metadata failure, not other forms of storage damage such as missing directories. Mike Degatano reviewed and approved the work.