If Performance Objectives are not being met, Quality of Service Management makes a recommendation. Each recommendation focuses on improving the highest-ranked Performance Class by exceeding its Performance Objective. Submissions may include changing consumer group mappings – and reprioritizing work within existing resource boundaries. For example, changing consumer group mappings may involve promoting a specific workload to get a more significant share of resources, or demoting a competing workload to make additional resources available to the target Performance Class. Another recommendation is to move servers between server pools and reprioritize resources between them to meet workload demands; so effectively, taking a node out of one pool and adding it to another pool gives more resources to the Performance Class running in that pool. And another recommendation is moving CPUs between databases within a server pool – reprioritize CPU resources within existing server pool boundaries. And this is called instance caging, where the CPU count parameter is set to limit the amount of CPUs an instance can use on a node.

The Quality of Service Management recommendations to improve the performance of a particular Performance Class adds more of the bottleneck resource – such as CPU time – for that Performance Class, making the bottleneck resource available more quickly to work requests in the Performance Class. Adding more resources to a Performance Class that is not performing well means taking resources away from another Performance Class. The Performance Class where the resources are removed should be less business-critical than the one being helped. So overall, the reallocation of resources should be beneficial to the business. When generating recommendations, Quality of Service Management evaluates the impact of the system performance as a whole. For example, suppose the improvement for one Performance Class is rather tiny, but the adverse effects on another Performance Class are significant. In that case, Quality of Service Management might report that the performance gain is too small and not recommended. If there is more than one way to resolve the bottleneck, Quality of Service Management advises the best overall recommendation. It is invariable, such as the calculated impact on all the Performance Classes and the predicted disruption and settling time associated with the action. And using Oracle Enterprise Manager, you can view the current and the alternative recommendations. Performance data is sent to Oracle Enterprise Manager for display on the Quality of Service Management Dashboard and Performance History pages. By default, Oracle Database QoS Management does not automatically implement recommendations. Instead, it suggests improving performance, which the administrator must then implement by clicking the Implement button. From version 12.1.0.2, Quality of Service Management allows you to specify authorized automatic actions that it can implement without the intervention of an administrator.

You query V$INSTANCE_RECOVERY view and consistently receive an OPTIMAL_LOGFILE_SIZE value that is greater than the size of your smallest online redo log file. The OPTIMAL_LOGFILE_SIZE column of the V$INSTANCE_RECOVERY view can be used to determine the appropriate size for all of the online redo log files in your database. If the value of the OPTIMAL_LOGFILE_SIZE column is greater than the size of your smallest online redo log file, you should change the size of all online redo log files to be at least this value. In addition, the FAST_START_MTTR_TARGET initialization parameter simplifies the configuration of recovery time from instance or system failure. After adjusting the size of your online redo log files, you may be able to adjust the value of this initialization parameter for better performance. This is done by rerunning the MTTR advisor after changing the size of your online redo log file to achieve more optimal results. However, running the MTTR advisor is not the best option in this situation for improving instance recovery performance.

 

You use the DBMS_RESOURCE_MANAGER package to create a CDB resource plan and define the directives for the plan. Then, from the root container of your CDB connects as the SYS user. Then, create a pending area using the CREATE_PENDING_AREA procedure. After the pending area has been completed, you use the CREATE_CDB_PLAN procedure to create the CDB resource plan. Next, create the CDB resource plan directives for the PDBs using the CREATE_CDB_PLAN_DIRECTIVE procedure. Each directive specifies how resources are allocated to a specific PDB. Finally, you validate the pending area and then submit it. This is done using the VALIDATE_PENDING_AREA and SUBMIT_PENDING_AREA procedures, respectively.