Database post 12.2 upgrade issue due to Scheduler Autotask running duplicate jobs during maintenance windows and increased resource consumption is observed:
According to MOS Doc ID 2387466.1 this is an internal bug (Bug 26724511 )
SYMPTOMS
After upgrading our database from 12.1 to 12.2 we found out that, some of the internal automated task queries are executed over and over by consuming high-CPU.
Actually, during the maintenance window, there is a special resource manager plan is in effect called ‘DEFAULT MAINTENANCE PLAN’ which prevents the autotask queries and DMLs to consume all the resource and limits them. This limitation is observed as a “resmgr: cpu quantum” wait event.
To summarize, after the upgrade, the autotask queries (statistics, space advisor, SQL tuning) started to be executed multiple times and require more CPU resources, causing the resource manager throttle them…
CAUSE:
Internal Bug (Bug 26724511 )
SOLUTION:
Apply the fix for the bug… Meanwhile, the gather stats and segment-advisor auto-tasks can be disabled until the fix is applied.
To disable the statistics auto-gathering:
SQL> exec dbms_scheduler.disable('GATHER_STATS_JOB');
To disable the space advisor:
SQL> begin DBMS_AUTO_TASK_ADMIN.disable(client_name=>'auto space advisor',operation=>NULL,window_name=>NULL);end;