is the goal to simulate concurrent load? if so, the proper way to do it is to gradually ramp up and then keep the load at a set level. Once the steady state is reached, take the desired measurements on the server.
from the description, it's not clear what the goal is. peter On 2/18/06, Tan, Kwei Fang <[EMAIL PROTECTED]> wrote: > > Hi, > > > > Need help and guideline on how to get the "started" time consistent for > 500 concurrent hits with the following setting: > > > > Thread Group Setting > > Number of Threads: 500 > > ramp-up period: 0 > > loop count:1 > > > > CSV Data Set config > > - set to read 500 lines of records for a txt file. > > > > Scenario: > > After run the jmeter test plan with above setting, all the 500 records > successfully process and displayed in jmeter.log > > > > Finding: > > As checking through the log file, noticed that the started time of > jmeter.threads seem to delay 1 second while reach some interval of > records, for example as following: > > ======================================================================== > =============================================== > > From line 1 to line 67: > > 2/16/2006 10:33:20 INFO - jmeter.threads.JMeterThread: > Thread Thread 11- 11350686442769 1-67 started > 11359080793880 10:36:30 > > > > At Line 68 to line 158: > > 2/16/2006 10:33:21 INFO - jmeter.threads.JMeterThread: > Thread Thread 11- 11350686442769 1-68 started > 11359080793897 10:36:31 > > > > At Line 159 to line 191: > > 2/16/2006 10:33:22 INFO - jmeter.threads.JMeterThread: > Thread Thread 11- 11350686442769 1-19 1 started > 11359080795129 10:39:36 > > ======================================================================== > =============================================== > > > > Question: > > The started time should be all the same from line 1 to line 500 records > as we are doing the concurrent thread running, anybody have any idea how > to make the "started" become consistent through the process? > > > > Thanks a lot. > > > > >
