Probably you seems to be losing connnections. Chunked response needs connections to be kept alive.
How many threads are you running and how many are erroring out ? With Regards, Kiran Badi Email:[email protected] Ph- US-(+01)6462013101 ________________________________ From: umesh prajapati <[email protected]> To: JMeter Users List <[email protected]> Sent: Thursday, July 25, 2013 11:12 PM Subject: Re: Error running the Test The reason it was sending different request was cookies were not cleared. After I checked mark the clear cookie in the cookie header. I ran the test again. and compared the request (of the user who successfully login on the first loop and were not able to login in the 2nd loop) I was not able to find any difference on the request. But I am still getting this error on my server log. ERROR[24-Jul-13 12:34:57] [http-0.0.0.0-9085-4] [ChunkedInputStream.parseTrailerHeaders()] - Error parsing trailer headers org.apache.commons.httpclient.ProtocolException: Unable to parse header: HTTP/1.1 200 OK at org.apache.commons.httpclient.HttpParser.parseHeaders(HttpParser.java:202) at org.apache.commons.httpclient.ChunkedInputStream.parseTrailerHeaders(ChunkedInputStream.java:322) at org.apache.commons.httpclient.ChunkedInputStream.nextChunk(ChunkedInputStream.java:226) On Thu, Jul 25, 2013 at 8:09 PM, umesh prajapati <[email protected]>wrote: > After I checked mark the clear cookie in the cookie header. I ran the test > again. and copied the request of the user who successfully login on the > first loop and were not able to login in the 2nd loop. > > > On Thu, Jul 25, 2013 at 4:52 PM, sebb <[email protected]> wrote: > >> On 25 July 2013 23:29, umesh prajapati <[email protected]> wrote: >> > Sorry I was wrong about saying that the request was identical..I >> realized >> > the request being send was different. >> >> What is the difference? >> >> > But I am wondering why is it sending different request for same user? >> >> A bug in the test script? >> >> > >> > On Thu, Jul 25, 2013 at 3:19 PM, umesh prajapati <[email protected] >> >wrote: >> > >> >> number of users: 50 >> >> ramp up period: 90 >> >> loop count: 5 >> >> >> >> So when I just ran this, on the view result tree I see that 3 users >> were >> >> not able to login the first time(first loop) but during the 2-5 loops >> they >> >> were successfully able to login. >> >> >> >> >> >> >> >> On Thu, Jul 25, 2013 at 12:42 PM, sebb <[email protected]> wrote: >> >> >> >>> On 25 July 2013 17:32, umesh prajapati <[email protected]> wrote: >> >>> > No difference on the request tab...I will go ahead and try changing >> the >> >>> > jmeter property and give it a try >> >>> >> >>> If the requests are identical, then it is starting to look like a >> >>> load-related issue - possibly a threading issue - at the servr. >> >>> >> >>> How many threads are you running? >> >>> >> >>> > >> >>> > On Thu, Jul 25, 2013 at 9:25 AM, sebb <[email protected]> wrote: >> >>> > >> >>> >> On 25 July 2013 17:02, umesh prajapati <[email protected]> >> wrote: >> >>> >> > jmeter logs: >> >>> >> > 2013/07/25 08:59:17 INFO - >> >>> >> > org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl$4: I/O >> exception >> >>> >> > (java.net.SocketException) caught when processing request: >> Connection >> >>> >> reset >> >>> >> > 2013/07/25 08:59:17 INFO - >> >>> >> > org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl$4: Retrying >> >>> request >> >>> >> > 2013/07/25 08:59:20 INFO - >> >>> >> > org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl$4: I/O >> exception >> >>> >> > (java.net.SocketException) caught when processing request: >> Connection >> >>> >> reset >> >>> >> > 2013/07/25 08:59:20 INFO - >> >>> >> > org.apache.jmeter.protocol.http.sampler.HTTPHC4Impl$4: Retrying >> >>> request >> >>> >> >> >>> >> Try setting the property >> >>> >> >> >>> >> httpclient4.retrycount=0 >> >>> >> >> >>> >> in jmeter.properties >> >>> >> >> >>> >> (will be the default for JMeter 2.10) >> >>> >> >> >>> >> This will report the error on the first sample, which is generally >> >>> >> what you want for performance testing. >> >>> >> Automatic retries can add unexpected traffic. >> >>> >> >> >>> >> > >> >>> >> > On Thu, Jul 25, 2013 at 8:47 AM, umesh prajapati < >> >>> [email protected] >> >>> >> >wrote: >> >>> >> > >> >>> >> >> I am looking at the view result tree (sample result tab)...The >> >>> users who >> >>> >> >> were able to login has load time (1000-3000) but the users who >> were >> >>> not >> >>> >> >> able to login has load time (60467-303469) >> >>> >> >> >> >>> >> >> >> >>> >> >> On Thu, Jul 25, 2013 at 8:43 AM, umesh prajapati < >> >>> [email protected] >> >>> >> >wrote: >> >>> >> >> >> >>> >> >>> I ran the test again using the http client 4 and I also checked >> >>> marked >> >>> >> >>> the browser compatible header but is still having the same >> issue >> >>> >> (login >> >>> >> >>> failures for some users) >> >>> >> >>> >> >>> >> >>> >> >>> >> >>> On Thu, Jul 25, 2013 at 8:32 AM, umesh prajapati < >> >>> [email protected] >> >>> >> >wrote: >> >>> >> >>> >> >>> >> >>>> I will let you know the result after I run the test using >> >>> httpclient4 >> >>> >> >>>> >> >>> >> >>>> >> >>> >> >>>> On Thu, Jul 25, 2013 at 8:31 AM, umesh prajapati < >> >>> >> [email protected]>wrote: >> >>> >> >>>> >> >>> >> >>>>> I just ran the test again and at the beginning of the jmeter >> log >> >>> >> before >> >>> >> >>>>> the test started I saw these but not really sure if this is >> >>> causing >> >>> >> not to >> >>> >> >>>>> login successful for some users >> >>> >> >>>>> >> >>> >> >>>>> 2013/07/25 08:11:00 INFO - >> >>> >> >>>>> jmeter.protocol.http.sampler.HTTPSamplerBase: Cannot find >> >>> .className >> >>> >> >>>>> property for htmlParser, using default >> >>> >> >>>>> >> >>> >> >>>>> 2013/07/25 08:14:20 INFO - jmeter.util.SSLManager: KeyStore >> >>> created >> >>> >> OK >> >>> >> >>>>> 2013/07/25 08:14:20 WARN - jmeter.util.SSLManager: Keystore >> >>> file not >> >>> >> >>>>> found, loading empty keystore >> >>> >> >>>>> >> >>> >> >>>>> >> >>> >> >>>>> >> >>> >> >>>>> On Thu, Jul 25, 2013 at 8:28 AM, sebb <[email protected]> >> wrote: >> >>> >> >>>>> >> >>> >> >>>>>> On 25 July 2013 16:12, umesh prajapati < >> [email protected]> >> >>> >> wrote: >> >>> >> >>>>>> > On my Http Request Currently I have these properties >> >>> >> >>>>>> > >> >>> >> >>>>>> > Implementation: Java >> >>> >> >>>>>> >> >>> >> >>>>>> Try using HttpClient 4 instead; >> >>> >> >>>>>> >> >>> >> >>>>>> The Java implementation is not ideal for testing. >> >>> >> >>>>>> >> >>> >> >>>>>> > Protocol: https >> >>> >> >>>>>> > Method: Get >> >>> >> >>>>>> > >> >>> >> >>>>>> > >> >>> >> >>>>>> > On Thu, Jul 25, 2013 at 8:10 AM, umesh prajapati < >> >>> >> >>>>>> [email protected]>wrote: >> >>> >> >>>>>> > >> >>> >> >>>>>> >> server logs >> >>> >> >>>>>> >> >> >>> >> >>>>>> >> I am using jmeter version 2.9 >> >>> >> >>>>>> >> >> >>> >> >>>>>> >> Can you try using a different HTTP implementation (i.e. >> not >> >>> >> Commons >> >>> >> >>>>>> >> HttpClient)? >> >>> >> >>>>>> >> do you mean to try different sampler request? >> >>> >> >>>>>> >> >> >>> >> >>>>>> >> >> >>> >> >>>>>> >> >> >>> >> >>>>>> >> On Thu, Jul 25, 2013 at 7:44 AM, sebb <[email protected]> >> >>> wrote: >> >>> >> >>>>>> >> >> >>> >> >>>>>> >>> On 25 July 2013 14:41, umesh prajapati < >> >>> [email protected]> >> >>> >> >>>>>> wrote: >> >>> >> >>>>>> >>> > Thank you for the response guys. Yes, I do have >> >>> assertions. >> >>> >> And >> >>> >> >>>>>> all the >> >>> >> >>>>>> >>> > users who were unable to login had those header >> response >> >>> and >> >>> >> all >> >>> >> >>>>>> the >> >>> >> >>>>>> >>> users >> >>> >> >>>>>> >>> > who were succesfully able to login had the other >> response >> >>> >> >>>>>> header. I ran >> >>> >> >>>>>> >>> th >> >>> >> >>>>>> >>> > test for 106 users to test the sso logins. I had the >> loop >> >>> for >> >>> >> >>>>>> two times. >> >>> >> >>>>>> >>> > >> >>> >> >>>>>> >>> > The first time I ran the test all the users were not >> >>> >> >>>>>> successfully able >> >>> >> >>>>>> >>> to >> >>> >> >>>>>> >>> > login. I tried running the test again and all the >> users >> >>> were >> >>> >> >>>>>> able to >> >>> >> >>>>>> >>> login >> >>> >> >>>>>> >>> > successfully. But I am getting this inconsistently. >> >>> >> >>>>>> >>> > >> >>> >> >>>>>> >>> > The error on the logs (not jmeter logs) >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >>> Which logs are they? >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >>> > ERROR[24-Jul-13 12:34:57] [http-0.0.0.0-9085-4] >> >>> >> >>>>>> >>> > [ChunkedInputStream.parseTrailerHeaders()] >> >>> >> >>>>>> >>> > - Error parsing trailer headers >> >>> >> >>>>>> >>> > org.apache.commons.httpclient.ProtocolException: >> Unable >> >>> to >> >>> >> >>>>>> parse >> >>> >> >>>>>> >>> header: >> >>> >> >>>>>> >>> > HTTP/1.1 200 OK >> >>> >> >>>>>> >>> > at >> >>> >> >>>>>> >>> > >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >> >>> >> >> >>> >> org.apache.commons.httpclient.HttpParser.parseHeaders(HttpParser.java:202) >> >>> >> >>>>>> >>> > at >> >>> >> >>>>>> >>> > >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >> >>> >> >> >>> >> org.apache.commons.httpclient.ChunkedInputStream.parseTrailerHeaders(ChunkedInputStream.java:322) >> >>> >> >>>>>> >>> > at >> >>> >> >>>>>> >>> > >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >> >>> >> >> >>> >> org.apache.commons.httpclient.ChunkedInputStream.nextChunk(ChunkedInputStream.java:226) >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >>> That looks like there might be an issue with the server >> >>> sending >> >>> >> bad >> >>> >> >>>>>> >>> data or the HttpClient parsing code not recognising the >> >>> >> response. >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >>> Which version of JMeter are you using? >> >>> >> >>>>>> >>> Can you try using a different HTTP implementation (i.e. >> not >> >>> >> Commons >> >>> >> >>>>>> >>> HttpClient)? >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >>> >> >>> >> >>>>>> >> >>> >> >> --------------------------------------------------------------------- >> >>> >> >>>>>> >>> 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] >> >>> >> >>>>>> >> >>> >> >>>>>> >> >>> >> >>>>> >> >>> >> >>>> >> >>> >> >>> >> >>> >> >> >> >>> >> >> >>> >> >> --------------------------------------------------------------------- >> >>> >> 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] >> >>> >> >>> >> >> >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> >> >
