Thanks for the replies. Let me first clear this out: the main purpose of my test is to generate load on our server. Preferably exactly the same load as during the stress test: the same request order at the same pace.
We monitor performance on the server, not in JMeter, so there is no listener added to the test. That is way the response is of little intrest to us. As for the users: these are request from many users (approx. 600 different users). I use a very simple custom CookieManager that returns the correct session ID cookie for each HTTP request, based on the user name that is logged in the HTTP logs. Splitting up the HTTP log per user looks like a good idea. The ordering per user is more important than ordering between users (but still can cause problems). Each per-user log will then be replayed by one JMeter thread. Now I can live with the fact that a user/thread waits for the response since now I have true concurreny between users. And with my custom Timer I can schedule users/threads based on the timestamp logged with each request. This comes close to what I want. Tnx for the replies! I'll give it a try next Monday (about to leave on short a vacation). Regards, Gerrie -- View this message in context: http://old.nabble.com/Replay-a-recorded-load-using-JMeter-tp28533968p28534808.html Sent from the JMeter - User mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
