Hello Don, > long startTime = System.currentTimeMillis(); > client.executeMethod(method); // method is a PostMethod > long elapsed = System.currentTimeMillis() - startTime;
I don't see what should be wrong with this approach. > > But what I'm seeing is that the threads all start up and mark the start > time, then the threads appear to be interrupted during executeMethod. > The result is that the measured times get progressively longer, with the > last thread to finish reporting a time almost as long as the entire > application's run time. If threads get interrupted, you'd see InterruptedException traces somewhere. If threads get blocked, you have to look for the reason of the blocking. For example, if all your requests go to the same host, then the default value for "max connections per host" will be too low. All threads will take turns in using (I think) two connections. >From what you describe, I would say that your measuring is correct and the bad response times you determine are the result of a badly configured connection manager. If the response times of the last threads are almost the runtime of the whole application, I suspect that each thread only executes one request. You should consider to have each thread execute several requests. cheers, Roland --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
