Red Hat Directory Server 7.1 Performance Tuning and Sizing Guidelines
13.
14.Schema checking: on (default) leave it on because it specifies whether the database schema will be enforced during entry insertion or modification. When measuring performance data based on exact searches, this core server configuration attribute won’t affect the performance. So leave it alone as the default value.
Selected Benchmark
DirectoryMark is selected as a benchmark for measuring performance of Red Hat Directory Server version 7.1. All the test data in this document is collected base on the DirectoryMark version 1.0. The latest version is 1.3 at the time of this writing, but it is only available on Windows 2000, 2003 servers, and Sun Solaris.
More information is available at http://www.mindcraft.com/directorymark/.
Test Machines
All the performance data are collected on the following machines:
Client:
Machine Class | Machine Specification |
SD32A | Within partition, |
Server: |
|
Machine Class | Machine Specification |
SD64B | Within the partition one cell with 4 |
| counted as 8 1.6GHz CPUs, 64GB memory, OS B.11.23 |
List 1: Client and Sever machine specifications.
Private LAN Configuration
All the test machines used for performance data collection are directly connected with Ethernet cables running point to point between cards. All the tests are done assuming there are no heavy network traffics on the private LAN. The configuration is as following:
Ethernet cables
SD32A server (client machine) (Itanium 2 processor)
Performance Monitoring Tools
vsar(Visual System Activity Reporter)
Vsar is an unsupported performance tool that periodically samples performance information from the kernel and displays that information on a user’s terminal. For more information about vsar, please refer to: http://webperf.cup.hp.com/perf_tools/vsarman.txt.
Page 16