This issue was resolved and the server is running correctly now. The OP's deployment was getting unintentionally "downgraded" to an earlier LabKey Server version by an automated nightly process. This explains the schema inconsistencies that were appearing every morning.