I just happened upon this thread.
Preface: I am on the Linux / server side of this equation .
Looking at the logs provided and reading about database corruption, not to be harsh, but points to the underlying system (either abrupt power-off or forced termination). Those two will corrupt a database in a blink every time. When running a server, a UPS, even a small one, will provide the time needed (only a few seconds) for PMS to shutdown gracefully and close the DB versus leaving stuff “scattered everywhere”.
As for the “Recommended” tab - that’s maintained by PMS in the database.
The very first thing I suspect is damage. Not necessarily DB corruption but a loss of sync between the DB and the cache files supporting it.
Something else I’m also seeing in the logs is a lot of media being analyzed.
There is no indication if this is new media or existing being re-analyzed for loudness, etc.
Either way, any media analysis will put a serious load on a system like this (performance level = 1700 Passmarks). My Recommendation is to go to Scheduled Tasks and look for things which don’t need constant updating and turn it off.
(e.g. Posters don’t update very often so why refresh them every 7 days ?)
