Cluster parameters

Last modified by Aurelie Bertrand on 2025/11/14 11:59

This section enables you to manage quotas and configure a cluster.

Quota management

It is possible to define different types of cube quotas applicable at different levels in order to limit the size of cubes or the memory used by cubes.

đź’ˇ It is possible to define cube quotas specific to data models (number of rows and size in memory only).
These quotas are stored in the data model itself and take precedence over those defined on this page.
See the page Define advanced cube parameters page for more details.

To do this:

  1. In the Quota field, select the type of quota you wish to add:
    CUBE_MAX_MEM_SIZEMaximum size of the cube in memory.
    Cube generation fails if the cube size exceeds that defined.
    CUBE_MAX_ROWSMaximum number of cube rows.
    Cube generation fails if the defined number of rows is exceeded.
    CUBE_MAX_MEM_USAGEMaximum memory used for loading cubes.
    Cube loading fails if the memory used by the cubes exceeds the defined memory.
  1. In the Quota scope field, select the level to which the quota applies:
    • GLOBAL: to all cubes (all roles and users)
    • ROLE: to cubes of the role selected in the drop-down list below.
    • PERSONAL: to personal portfolios. The quota value applies to all personal portfolios (all users).
  2. In the Quota limit field, enter the quota limit value as an absolute value or as a percentage.
    Specify the unit following the value in the next field:
    • Maximum size or memory: K=Kilo byte, M=Mega byte, G=Giga byte and %=percentage.
    • Maximum number of lines: k=thousand, m=million.
  3. Click on the Add button.
    ➡ The quota is added to the Quota List field.

đź’ˇYou can define the same quota several times with a different scope. Quotas are applied in order of priority PERSONAL -> ROLE -> GLOBAL.

Quotas

Examples

Example 1

A quota of type CUBE_MAX_MEM_USAGE is defined with a GLOBAL scope and a limit of 30%.
Another quota of type CUBE_MAX_MEM_USAGE is defined with a ROLE scope on the Retail role and a limit of 50%.
In this case, the memory used for loading all the cubes cannot exceed 30% of the memory. Loading a cube will fail if this limit is exceeded.
On the other hand, the memory used to load the Retail role cubes can be up to 50% of memory. The ROLE scope quota takes precedence over the GLOBAL scope quota.

Example 2

A quota of type CUBE_MAX_ROWS is defined with a GLOBAL scope and a limit of 500k (500,000 lines).
Another quota of type CUBE_MAX_ROWS is defined with a ROLE scope on the Telecom role and a limit of 1m (1 million lines).
In this case, the generation of a cube for the Telecom role will fail if its number of lines exceeds 1 million, whereas the generation of a cube for any other role will fail if the number of lines in the cube exceeds 500,000.

Configuring a cluster

To manage a larger volume of data (one billion rows), it is possible to use several servers in "Cluster" mode.

In the case of internal clustering, you can configure the cluster here. See the paragraph Configuring the cluster section on the Using several servers in "Cluster" mode page for more details.