On 13/03/2009, Abel MacAdam <[email protected]> wrote: > > What are you trying to prove? > > I was asked to tell what the load was the server could deliver. Or stated in > other words, how many concurrent users the server could handle. >
In that case, you probably need to run several tests, with increasing numbers of threads. See when the server starts to struggle. Make sure that the network connectivity is adequate or that could be the limiting factor. I would start with a single thread, download the file and save it using "Save Response to a file". Check that the file plays OK locally, and get its MD5. Try again, using save Response as MD5 and add an Assertion to check the MD5 is OK. Then try increasing the thread count until things go wrong. > > Start a thread, and add an assertion that the film completes within the > > allotted time? How? > > Please read the manual: > > > http://jakarta.apache.org/jmeter/usermanual/component_reference.html#Duration_Assertion > > That I did. I already have the following added to my script: > Duration Assertion: Duration in milliseconds 1525 (being 25 minutes and 25 > seconds, the duration of the movie). But am I measuring the correct thing??? That was not clear from your earlier question. > > > And how would I simulate several users at the same time? > > Please read the manual: > > > http://jakarta.apache.org/jmeter/usermanual/component_reference.html#Thread_Group > > > Sigh. That I also did, given that I simulate 1000 users with a startup time > of 10 seconds, repeating forever. In that case, I don't understand the question. > Thanks for your feedback. > > > > sebb-2-2 wrote: > > > > On 13/03/2009, Abel MacAdam <[email protected]> wrote: > >> > >> 1000 Threads in 10 seconds looping forever was a test I performed today. > >> Although it ran I'm having issues with the correctness of the test. So > >> therefore this question. > >> > >> So JMeter waits 'until the GET request has completed.' As a result, the > >> librarian started reading the first page. But will the librarian keep on > >> reading according to JMeter? > > > > This is not up to JMeter. JMeter reads data until the server tells it > > there is no more to come - or disconnects. > > > > I don't know what the server does. > > > >> If so, what would be a more appropriate test? > > > > What are you trying to prove? > > > >> Start a thread, and add an assertion that the film completes within the > >> allotted time? How? > > > > Please read the manual: > > > > > http://jakarta.apache.org/jmeter/usermanual/component_reference.html#Duration_Assertion > > > >> And how would I simulate several users at the same time? > > > > Please read the manual: > > > > > http://jakarta.apache.org/jmeter/usermanual/component_reference.html#Thread_Group > > > >> Is JMeter an appropriate system to test the load of a streaming media > >> server? > >> > >> Thanks. > >> > >> > >> > >> sebb-2-2 wrote: > >> > > >> > On 13/03/2009, Abel MacAdam <[email protected]> wrote: > >> >> > >> >> Hi, > >> >> > >> >> Can someone explain to me what happens when I open a page containing > >> >> streaming media (a flash movie) in JMeter? > >> >> > >> >> In a normal JMeter session I think you test something like (sort of) > >> a > >> >> (string of) request - response combination(s). So each time JMeter > >> goes > >> >> to > >> >> the library (the request), it fetches a book (the response). With > >> >> streaming > >> >> media however, you send JMeter to the library, then asks the > >> librarian > >> >> to be > >> >> read from the book. And if I am going to test what load the > >> streaming > >> >> media > >> >> server can support, I'm sending many persons concurrently requesting > >> to > >> >> be > >> >> read from the book. And to me, fetching a book is not so though as > >> >> asking to > >> >> be read from the book. > >> > > >> > JMeter will wait until the GET request has completed. > >> > I'm not sure if this behaves differently for streaming media. > >> > > >> >> Now back to JMeter. What if I have a thread group with 1000 threads, > >> to > >> >> be > >> >> started in 10 seconds, looping forever. > >> > > >> > 1000 threads may be too many for one system. > >> > > >> >> Back to my question: If I test a connection with a streaming media > >> >> server, > >> >> what do I do? Fetch 1000 books every 10 seconds, or request 1000 > >> >> librarians > >> >> to read me a book every 10 seconds? > >> > > >> > Again, what are you trying to prove? > >> > > >> >> Abel > >> >> > >> >> -- > >> >> View this message in context: > >> >> > >> > http://www.nabble.com/Follow-up-on-streaming-media-tp22496310p22496310.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] > >> >> > >> >> > >> > > >> > --------------------------------------------------------------------- > >> > To unsubscribe, e-mail: [email protected] > >> > For additional commands, e-mail: [email protected] > >> > > >> > > >> > > >> > >> > >> -- > >> View this message in context: > >> > http://www.nabble.com/Follow-up-on-streaming-media-tp22496310p22497232.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] > >> > >> > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [email protected] > > For additional commands, e-mail: [email protected] > > > > > > > > -- > > View this message in context: > http://www.nabble.com/Follow-up-on-streaming-media-tp22496310p22498765.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] > > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
