Hitachi

Hitachi Advanced Database Setup and Operation Guide


4.3 Estimating memory requirements

This section explains how to estimate the memory requirements based on the values obtained in 4.2 Estimating resource requirements. The memory used by the HADB server consists of shared memory and process memory. There are four types of shared memory:

If you wish to estimate the approximate memory requirements before designing a database, use the formulas described in the subsections below. Note, however, that the values obtained are only rough estimates, since it is impossible to estimate memory requirements in any great detail.

The amount of memory used by the HADB server differs greatly depending on which of the following processing is performed.

■ Processing executed after the HADB server starts
  • Execution of normal operations (database connection, SQL statement execution)

  • Execution of the adbimport command to import data

  • Execution of the adbidxrebuild command to rebuild indexes

  • Execution of the adbgetcst command to collect cost information

  • Execution of the adbexport command to export data

  • Execution of the adbmergechunk command to merge multiple chunks

  • Execution of the adbarchivechunk command to archive a chunk

  • Execution of the adbunarchivechunk command to unarchive a chunk

We recommend that you determine the amount of memory that will be used for each of these types of processing, and then use the largest of the values that are obtained as the amount of memory the HADB server will use. For details about performing normal operations (database connection, SQL statement execution):

If you want to estimate memory requirements more precisely, see 6.3 Estimating the HADB server's memory requirement after you have designed the database and estimated the DB area size.

Organization of this section