224 IBM Cognos Dynamic Cubes
aggregate cache load is to have supporting database aggregates. These database
aggregates can be created by the DBA, based on specific needs.
To help determine which supporting database aggregates can be used to increase the speed
of aggregate cache load, view the information in the database recommendation output. if the
Aggregate Advisor was run to include both in-memory and in-database recommendations,
information in the in-database recommendation textual output indicates how many
corresponding in-memory aggregate recommendations can be satisfied by the database
aggregate recommendation. For example, the header section for a logical database
aggregate might contains a line similar to the following line:
Number of recommended in-memory aggregates that this aggregate can cover: 5
Controlling the number of aggregates loaded in parallel
By default, the number of aggregates that are loaded concurrently is determined as being
twice the number of processors of the QueryService system. For example, if the
QueryService system has four processors, then there will be eight threads to load the
aggregates in parallel.
The number of queries posed concurrently to populate the in-memory aggregate cache can
be controlled by an advanced property of Cognos Dynamic Cubes to ensure the underlying
relational database is not saturated with concurrent requests computing summary values.
To reduce the number of aggregate load queries against the underlying database at one time,
reduce the value for the qsMaxAggregateLoadThreads property. Note that fewer threads more
require more overall time to load the aggregates.
1. Go to IBM Cognos Administration.
2. Select Status.
3. Select System.
4. Select the server.
5. Select the dispatcher.
6. Select QueryService.
7. In the QueryService context menu, select Set properties.
8. Select the Settings tab.
9. In Advanced settings, select Edit in the Value column.
10.Select the Override the settings from the parent entry check box.
11.In an available row, enter qsMaxAggregateLoadThreads under the Parameter field, and a
value under the corresponding Value field.
12.Click OK.
The setting is a global setting for the QueryService, where each cube has the specified
number of threads to load aggregates in parallel.
Note: The DBA should be aware of the impact to the underlying relational database
resources that the aggregate cache load activities can have. So, the DBA should be aware
of the cube administration tasks that result in an aggregate cache load, such as cube start,
restart, data cache refresh, and member cache refresh.