How do I fix a Kaspa node stuck after a database error?

Navigate to %localappdata%\Kaspad\kaspa-mainnet\datadir2 and delete the newest few files, then restart your node — repeat as needed. Kaspad stores its blockchain data in a local folder called datadir2; a mid-crash write can leave corrupted files at the end of that folder, and removing them lets kaspad resume from a clean checkpoint. If deleting a few files doesn't help, delete everything in the folder to resync from scratch, or launch kaspad with the --reset-db command-line parameter to let it wipe the database itself — but remove that flag afterward, or kaspad will wipe the database every single time it restarts. Working through these steps in order resolves most node crashes without needing outside help.