jmeter-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From bugzi...@apache.org
Subject [Bug 62045] timing difference for large (?) requests
Date Wed, 31 Jan 2018 17:02:50 GMT
https://bz.apache.org/bugzilla/show_bug.cgi?id=62045

--- Comment #7 from Philippe Mouawad <p.mouawad@ubik-ingenierie.com> ---
(In reply to Alex Podelko from comment #6)
> Even in JMeter it takes 37 sec - while in real life it takes 75 sec. There
> are few second variations, but overall the ratio is rather stable.
> 
> Even if forget about time to do it repeatable - it is a part of the process
> and it creates a large number of entries. Doing that many times in a loop
> will create other performance issues in the system.
> 
> That large request is specified more for example - because the difference is
> so large and evident that you can't ignore / misinterpret it. 

Ok, but in my experience, I have faced applications that were very unstable in
terms of response time for the same request (not under load, but really during
scripting/recording) but ran at different timings.
That's why I was asking for this particularly because the logs you attach show
that you ran them at very different timings:
- 15h08
- 13h28


Can you attach your jmeter.log and the gc.log after uncommenting in jmeter.bat
this part:
set VERBOSE_GC=-verbose:gc -Xloggc:gc_jmeter_%%p.log -XX:+PrintGCDetails
-XX:+PrintGCCause -XX:+PrintTenuringDistribution -XX:+PrintHeapAtGC
-XX:+PrintGCApplicationConcurrentTime -XX:+PrintGCApplicationStoppedTime
-XX:+PrintGCDateStamps -XX:+PrintAdaptiveSizePolicy

> 
> In most cases I use smaller requests in tests - where we still have the
> difference, but it is smaller.


Frankly I have no idea of what is happening and it is very difficult to figure
out without have the application available.

-- 
You are receiving this mail because:
You are the assignee for the bug.
Mime
View raw message