Following an upgrade to DX UIM 23.4.6 (CU6), the following issues are observed:
wasp.log or OC logs show: ERROR com.firehunter.operatorconsole.service.ComputerSystemService:searchComputerSystems - Issue in capturing Maintiance State java.lang.NullPointerException: Cannot invoke "java.util.Map.size()" because "activeScheduleList" is null.mon_config_service.log may grow significantly (3GB+) and become locked by java.exe processes.
When the OC Inventory loads up, it also try to pull maintenance states of devices. An accumulation of records and/or a deadlock in the MAINTENANCE_WINDOW_HISTORY table prevents the ComputerSystemService from correctly processing the maintenance state for inventory devices, leading to a NullPointerException and API timeouts.
The MAINTENANCE_WINDOW_HISTORY and related maintenance tables lack necessary indexes by default in certain 23.4.x versions. As these tables grow, the maintenance_mode probe and wasp experience severe performance degradation and deadlocks, preventing the ComputerSystemService from processing maintenance states.
Note: If you stop mainteinance_mode, the inventory loads immediately.
maintenance_mode and wasp probes.TRUNCATE TABLE MAINTENANCE_WINDOW_HISTORY;Deploy the maintenance_mode-23.4.6-T1.zip (or later) hotfix to your primary UIM Server. This version includes the updated table indexes and fixes the issue causing blocked updates.
To prevent recurrence, configure auto-purging and regular database maintenance:
maintenance_mode probe configuration, set: purge_windows_history_days = 30ALTER INDEX ALL ON [dbo].[MAINTENANCE_WINDOW] REBUILD WITH (ONLINE = ON);ALTER INDEX ALL ON [dbo].[MAINTENANCE_WINDOW_HISTORY] REBUILD WITH (ONLINE = ON);ALTER INDEX ALL ON [dbo].[MAINTENANCE_SCHEDULE] REBUILD WITH (ONLINE = ON);ALTER INDEX ALL ON [dbo].[MAINTENANCE_SCHEDULE_MEMBERS] REBUILD WITH (ONLINE = ON);NOTE: For long-term prevention Broadcom is enhancing the maintenance_mode probe ability to auto-purge very old audit information.