| 摘要 |
<p>A method of benchmark testing a component 112 of a first computer system (e.g. a server) 103 such as a database system. The method uses test data 109 and an additional computer system (server) 102 to restore the data to an initial state 110 so that the benchmark may be repeated from an identical starting point. Another component of the system can then be tested. The component tested may include the application program, the memory, CPU speed or the number of CPUs. The method may include the steps of creating a master copy of the test data 109; linking the additional computer 102 to additional storage 110 by switching using a logical volume controller; copying the test data 109 to the additional storage 110; severing the link from the additional computer 102 to the additional storage 110; linking the additional storage 110 to the first computer 103 and executing a benchmark test using the first computer system 103. The additional computer may then make a further copy of the master test data. This copy may then be cycled with the data previously used to effect a rollback of the test data allowing further benchmark testing.</p> |
