As a long-time user of Deconz, I’ve recently encountered a significant issue with my database growing uncontrollably to over 800MB. This has led to extremely slow startup times for the Deconz service, often taking over 20 minutes to initialize. While I’ve managed to keep everything updated, including the Raspberry Pi 4, RaspBee II, and the latest Deconz version, the database corruption has become a major hurdle.
I’ve tried several troubleshooting steps, including backing up and restoring the database through the Phoscon web interface, but this resulted in the loss of half my devices. Using the DB Browser for SQLite to manually delete entries also proved fruitless, as the application wouldn’t allow row deletions. I came across a similar issue on GitHub, but the suggested solution didn’t work for me due to numerous UNIQUE constraint failures.
This problem is quite concerning, especially since my setup includes around 80 Aqara devices. The database corruption is clear, but I’m at a loss on how to clean it up without losing all my sensors and switches. If anyone has successfully resolved this issue or has insights into effective database optimization techniques, I’d greatly appreciate your input. Let’s work together to find a solution!