Followed your instructions the JSESSION ID's are unique and they are being sent correctly. What tipped you off that something is wrong with the log in processing? Your thoughts are appreciated... I'll speak with dev..
Thanks. Mordechai sebb-2-2 wrote: > > On 25/02/2009, Mordechai <mr.happy.sna...@gmail.com> wrote: >> >> It seems like Jmeter _maybe_ messing up the jsessions when the users are >> launched to closely together. >> >> If you have 2 Thread Groups one called A another called B and each >> contains >> 5 users and they both launch 5 users in 1 second. That's 10 users >> launching in 1 second from 2 Thread Groups, most users error out meaning >> they do not log in to Information Server successfully. >> >> If you make Thread Group A with 5 users have a ramp up of 5 seconds >> meaning >> 1 thread / per second is launched then Thread Group B with 5 users have >> a >> ramp up of 10 seconds meaning 1 thread / per 2 seconds is launched >> there is >> only 1 user who errors out. Thread Group B user 1 or the >> 6tth thread to be launched does not log in correctly. >> >> And here are the exact times of log ins >> >> USER_id Time of log in attempt >> perf 1, 2009-02-11 13:34:51 >> perf 2, 2009-02-11 13:34:52 >> perf_3, 2009-02-11 13:34:53 >> perf_4, 2009-02-11 13:34:54 >> perf_5, 2009-02-11 13:34:55 >> >> perf 11, 2009-02-11 13:34:51 * failed * >> perf 12, 2009-02-11 13:34:53 >> perf 13 , 2009-02-11 13:34:55 >> perf 14, 2009-02-11 13:34:57 >> perf 15, 2009-02-11 13:34:59 >> >> >> What breaks my little theory is that perf_3 and perf_12 both should >> also >> error out but they do not. Perhaps they difference was slighter than a >> second...either way if I run these groups into a loop they will start >> error-ing out as User Perf 11 has. >> >> I've been using Jmeter for several months and have never run into an >> issue >> like this. The only difference is that this application uses a cookie >> to >> store its jsession. >> >> I ran the test several times with different configuration in terms of >> ramp >> up time and number of users. When ever two threads try to log in at the >> same time from the same client machine only one of the users logs in, >> though >> once's he's logged in he's fine until he has to log in again. > > That sounds more like a problem with the login processing on the server. > > But in order to prove this, you will have to track the values of the > cookies. Since cookies are saved in JMeter variables, you can add the > cookie names to the JMeter property sample_variables in order to have > the values written to the log file: > > http://jakarta.apache.org/jmeter/usermanual/listeners.html#sample_variables > > Use CSV format for the log file (smaller file, less resource hungry than > XML). > >> Jmeter manual mentions that cookies with the same name or written over >> could >> this be causing this? >> > > No; this is normal for cookies. Only the last value for a cookie is > stored, otherwise servers could never update cookie values. > >> >> >> sebb-2-2 wrote: >> > >> > On 24/02/2009, Mordechai <mr.happy.sna...@gmail.com> wrote: >> >> >> >> Sorry to ask a dumb question. But why is it significant not to >> re-use >> >> the >> >> same user name? >> > >> > Depends on the application. >> > >> > It may not like the same user logging in from two different sessions. >> > >> >> The csv's that I'm using contain seperate users, but the issue that >> I'm >> >> having occurs the same with 1 user name or with 10 different ones >> being >> >> used. >> >> >> > >> > In that case, it seems it's not the user name. >> > >> > But if the login details are correct, why is the login failing? >> > >> > Are you sure that the threads logout correctly? >> > >> >> >> >> sebb-2-2 wrote: >> >> > >> >> > On 24/02/2009, Mordechai <mr.happy.sna...@gmail.com> wrote: >> >> >> >> >> >> my answers below in bold >> >> >> >> >> > >> >> > Not here they're not ... >> >> > >> >> >> >> >> >> sebb-2-2 wrote: >> >> >> > >> >> >> > On 24/02/2009, Mordechai <mr.happy.sna...@gmail.com> wrote: >> >> >> >> >> >> >> >> Hi, >> >> >> >> >> >> >> >> After reading your posts I wanted to better describe the >> problem >> >> I'm >> >> >> >> having. >> >> >> >> >> >> >> >> Problem: When executing a test plan that uses a cookie to >> store >> >> the >> >> >> >> session >> >> >> >> id, >> >> >> >> the number of users who log into the test application >> >> successfully >> >> >> are >> >> >> >> not >> >> >> >> consistent. >> >> >> >> My Test plan executes correctly with 1- 5 threads. >> >> >> >> When more than 5 threads are used I notice that some of the >> >> users do >> >> >> not >> >> >> >> log >> >> >> >> in correctly. >> >> >> > >> >> >> > That could perhaps be due to a fault in the server, or an >> error in >> >> the >> >> >> > CSV data used to define the users. >> >> >> > >> >> >> >> Correctly means they are shown in Information Server console >> -> >> >> >> active >> >> >> >> sessions >> >> >> >> as an active connection. Correctly also means that these >> users >> >> do >> >> >> not >> >> >> >> fail >> >> >> >> against assertions >> >> >> >> that check for specific html in request response to assert >> that >> >> >> there >> >> >> >> logged >> >> >> >> in. >> >> >> >> >> >> >> >> When I check the View Results Tree, I am able to see that a >> user >> >> >> logged >> >> >> >> in >> >> >> >> correctly because >> >> >> >> I have assertions that check for expected html in a response >> >> like >> >> >> >> "Welcome". >> >> >> > >> >> >> > OK. >> >> >> > >> >> >> >> Application: >> >> >> >> A snippet of the first few requests are the following: >> >> >> >> -TestPlan >> >> >> >> --Threadgroup >> >> >> >> ---CSV >> >> >> > >> >> >> > What is the CSV element used for? >> >> >> > >> >> >> >> >> >> > The csv element is used to store user names. >> >> >> > Yes the correct usernames are being used. >> >> >> > Yes the user names have proper websphere permissions to log >> in. >> >> > >> >> > Are you sure you are not accidentally reusing a user name? >> >> > >> >> >> > >> >> >> >> ---UserParameter >> >> >> >> ---View Results Tree >> >> >> >> ---Http Cookie Manager >> >> >> >> ---Http Header Manger ( contains entry User-Agent: >> >> >> >> Mozilla/4.79[en](X11;U;Linux 2.2.12 i386) >> >> >> >> ---Aggregate Report >> >> >> >> +--Transactoin: Script Name >> >> >> >> +---Transaction: LogIn Page >> >> >> >> -----/appname >> >> >> >> -----/appname/ (Header Recieved Set-Cookie >> >> >> JSESSIONID=00PW8JoNQgW:-1; >> >> >> >> Path=/ >> >> >> >> -----/appname/searchform.do (Header sent with JSESSIONID) >> >> >> > >> >> >> > Does it send the correct cookie as received above? >> >> >> > >> >> >> >> >> >> > Thats exactly the issue, the only way to see if a cookie >> JSession >> >> >> cookie >> >> >> > value was sent >> >> >> > is through the ViewResultTree and clicking on the actual >> request. >> >> > >> >> > So? >> >> > >> >> >> > When you do so it shows a cookie value in the header sent with >> a >> >> >> jsession >> >> >> > id. >> >> >> > I'll check with fewer threads to see if its sending the right >> one. >> >> It >> >> >> > seems that >> >> >> > it's obviously sending the right ones most of the time, and >> not >> >> when >> >> >> the >> >> >> > users go >> >> >> > above 5 or 6 users. Is there anyway to verify this out >> outside of >> >> >> > matching up >> >> >> > requests and response in View Result? >> >> > >> >> > No. >> >> > >> >> >> > >> >> >> >> -----/appname/login.jsp >> >> >> >> -----/appname/j_acegi_security_check.do (user id / pw sent >> here >> >> ) >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> The cookie received in /appname/ I do not see in any headers >> in >> >> >> Jmeter. >> >> >> > >> >> >> > What do you mean? >> >> >> > This does not agree with "(Header sent with JSESSIONID)", nor >> with >> >> the >> >> >> > output from the http debugger shown below. >> >> >> > >> >> >> >> >> >> > I was saying that I expected to see in the Jmeter test plan >> >> recording >> >> >> a >> >> >> > value in the header that said >> >> >> > cookie: and a value that was the jsession id, as I do when I >> view >> >> the >> >> >> > request through Httpwatch. >> >> > >> >> > We've already established that cookies are not shown in the test >> plan. >> >> > This is because they vary between runs, unlike other headers. >> >> > >> >> >> > The line:with "(Header sent with JSESSIONID)" >> >> >> > was informing you that is when the cookie should be sent. I >> can >> >> show >> >> >> that >> >> >> > is sent only by viewing the View Results tree. >> >> >> >> >> >> > >> >> >> > >> >> >> >> As I had posted earlier I incorrectly expected to see it >> sent >> >> as a >> >> >> part >> >> >> >> of >> >> >> >> subsequent header requests. Using other http debuggers >> >> >> >> I can see that the next request contains the following >> header: >> >> >> >> >> >> >> >> >> >> >> >> (Request-Line) GET /appname/searchForm.do HTTP/1.1 >> >> >> >> Accept image/gif, image/x-xbitmap -shortened for this >> email- >> >> */* >> >> >> >> Accept-Encoding gzip, deflate >> >> >> >> Accept-Language en-us >> >> >> >> Connection Keep-Alive >> >> >> >> Cookie JSESSIONID=00PW8JoNQgW:-1 >> >> >> > >> >> >> > So the cookie is sent. Is the value the expected value? >> >> >> > >> >> >> >> Host pmachine-x:9080 >> >> >> >> User-Agent Mozilla/4.0 (compatible; MSIE 6.0; Windows >> NT >> >> 5.1; >> >> >> SV1; >> >> >> >> Avant >> >> >> >> Browser) >> >> >> >> >> >> >> >> The cookie jsession cookie is sent to the server in >> >> >> >> /appname/searchform.do >> >> >> > >> >> >> > Is it the correct value? >> >> >> > >> >> >> >> When debugging my script through View_Results Tree I can see >> >> >> >> that the requests are sending a jsession cookies. >> >> >> >> >> >> >> >> What am I dong wrong? When should I be using a header >> manager? >> >> >> > >> >> >> > When you need to send a header (other than a cookie). >> >> >> > >> >> >> >> My problem is very very similiar to thread: >> >> >> >> http://marc.info/?l=jmeter-user&m=100867164302881&w=2 >> >> >> >> except the fix did not work for me. >> >> >> > >> >> >> > I don't see the relevance myself. >> >> >> > >> >> >> > >> >> >> >> >> >> > The relevance is that he was trying to send a cookie that >> contained >> >> a >> >> >> > jsession id >> >> >> > and it wasnt working just as its not working for me. >> >> >> >> >> > >> >> > But as far as I can understand from this thread, the cookies are >> being >> >> > sent. >> >> > >> >> >> > >> >> >> > >> >> >> >> >> >> >> >> Pictures of test plan >> >> >> >> >> >> https://www.getdropbox.com/gallery/398664/1/Jmeter-Issues?h=2a7cab >> >> >> >> >> >> >> >> a) How do you appropriately interact with cookies in this >> >> situation? >> >> >> ( >> >> >> >> I'm >> >> >> >> using a cookie manager) >> >> >> >> b) What obvious mistake am I making? >> >> >> > >> >> >> > What does the server say on the page which normally has the >> >> welcome >> >> >> > message? >> >> >> > >> >> >> >> >> >> > WHat do you mean say? WHat is the full response? I dont >> >> understand. >> >> >> If >> >> >> > I log in >> >> >> > and things are fine the page displays nromal content. If I log >> in >> >> and >> >> >> its >> >> >> > one of the threads >> >> >> > that are not able to log in, the server returns an error page >> >> >> requiring >> >> >> > log in >> >> > >> >> > So what exactly does the error page say? >> >> > >> >> >> > >> >> >> > That may give a clue as to what is going on. >> >> >> > >> >> >> >> Thank you for your patience and help in advance. >> >> >> >> Mordechai >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> On 23/02/2009, Mordechai <mr.happy.sna...@gmail.com> wrote: >> >> >> >> > >> >> >> >> >> >> >> >> There's something wrong with the way previous messages are >> being >> >> >> >> quoted when you reply. >> >> >> >> >> >> >> >> For some reason an extra > is being added to your responses. >> >> This >> >> >> >> makes it very difficult to follow the thread. >> >> >> >> >> >> >> >> > >> >> >> >> > sebb-2-2 wrote: >> >> >> >> > > >> >> >> >> > > On 23/02/2009, Mordechai <mr.happy.sna...@gmail.com> >> wrote: >> >> >> >> > >> >> >> >> >> > >> >> >> >> >> > >> >> >> >> >> > >> sebb-2-2 wrote: >> >> >> >> > >> > >> >> >> >> > >> > On 23/02/2009, Mordechai >> <mr.happy.sna...@gmail.com> >> >> wrote: >> >> >> >> > >> >> >> >> >> >> > >> >> Hi there, >> >> >> >> > >> >> >> >> >> >> > >> >> Thanks for the reply. I am using the cookie >> manager. >> >> It >> >> >> >> seems >> >> >> >> > the >> >> >> >> > >> >> jsession >> >> >> >> > >> >> is sent in a cookie immediately before the log in >> >> page. >> >> >> >> > >> > >> >> >> >> > >> > Does the test plan include the page before the >> login >> >> page? >> >> >> >> > >> > >> >> >> >> > >> >> >> >> >> > >> > - Yes the 1st request contains the cookie >> JSESSIONID >> >> >> >> > >> >> >> >> >> >> >> >> >> >> >> >> >> -- >> >> >> >> View this message in context: >> >> >> >> >> >> >> >> >> >> http://www.nabble.com/Jmeter-proxy-recoring-not-showing-cookie-tp22150874p22184707.html >> >> >> >> >> >> >> >> Sent from the JMeter - User mailing list archive at >> Nabble.com. >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> --------------------------------------------------------------------- >> >> >> >> To unsubscribe, e-mail: >> >> jmeter-user-unsubscr...@jakarta.apache.org >> >> >> >> For additional commands, e-mail: >> >> jmeter-user-h...@jakarta.apache.org >> >> >> >> >> >> >> >> >> >> >> > >> >> >> > >> >> --------------------------------------------------------------------- >> >> >> > To unsubscribe, e-mail: >> jmeter-user-unsubscr...@jakarta.apache.org >> >> >> > For additional commands, e-mail: >> >> jmeter-user-h...@jakarta.apache.org >> >> >> > >> >> >> > >> >> >> > >> >> >> >> >> >> >> >> >> -- >> >> >> View this message in context: >> >> >> >> >> >> http://www.nabble.com/Jmeter-proxy-recoring-not-showing-cookie-tp22150874p22189023.html >> >> >> >> >> >> Sent from the JMeter - User mailing list archive at Nabble.com. >> >> >> >> >> >> >> >> >> >> >> --------------------------------------------------------------------- >> >> >> To unsubscribe, e-mail: >> jmeter-user-unsubscr...@jakarta.apache.org >> >> >> For additional commands, e-mail: >> jmeter-user-h...@jakarta.apache.org >> >> >> >> >> >> >> >> > >> >> > >> --------------------------------------------------------------------- >> >> > To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org >> >> > For additional commands, e-mail: >> jmeter-user-h...@jakarta.apache.org >> >> > >> >> > >> >> > >> >> >> >> -- >> >> >> >> View this message in context: >> >> >> http://www.nabble.com/Jmeter-proxy-recoring-not-showing-cookie-tp22150874p22190185.html >> >> >> >> Sent from the JMeter - User mailing list archive at Nabble.com. >> >> >> >> >> >> >> --------------------------------------------------------------------- >> >> To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org >> >> For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org >> >> >> >> >> > >> > --------------------------------------------------------------------- >> > To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org >> > For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org >> > >> > >> > >> >> -- >> >> View this message in context: >> http://www.nabble.com/Jmeter-proxy-recoring-not-showing-cookie-tp22150874p22201898.html >> >> Sent from the JMeter - User mailing list archive at Nabble.com. >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org >> For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org >> >> > > --------------------------------------------------------------------- > To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org > For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org > > > -- View this message in context: http://www.nabble.com/Jmeter-proxy-recoring-not-showing-cookie-tp22150874p22206564.html Sent from the JMeter - User mailing list archive at Nabble.com. --------------------------------------------------------------------- To unsubscribe, e-mail: jmeter-user-unsubscr...@jakarta.apache.org For additional commands, e-mail: jmeter-user-h...@jakarta.apache.org
