jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From William Ottley <>
Subject confused again about users/threads
Date Wed, 24 Mar 2010 17:20:59 GMT
I think I have been doing my tests wrongly, and not starting off with low
Now that I have decided to get back to basics,
reasons: (is it OS, or jmeter: open files, WAIT, etc)

I have a sampler that is setup as:
Thread Group: 500 threads, 0 ramp-up, and 5 Loops
HTTP Request Defaults (location and port)
1x HTTP Request HTTPClient: index.html

I did this, thinking that I would get errors, but I didn't... strange,
so I set it to 1000 threads, and then 2000 threads, then 3000 threads, and I
was like ok... this makes no sense.

I went to the web server side, and I only saw from 3 to 10 connections at a
I went and verified logs: all requests are coming back "OK"

So, now I have to read the results as this:
there is a single user, requesting 500 "index.html" files, 1 AFTER ANOTHER,
and NOT simultaneously...
am I correct with this logic?

If so, HOW does one go about then, simulating 500 users, requesting 500
"index.html" files?

does that mean I have to have 500 x "Thread Group"? and each of those Thread
Group will have 500 threads?

I suspect, I would, in normal situations have something like 5 or 10 of
these Thread Groups, and have different servers run the tests at the same

OR, I could have 500 different servers, all running a single Thread Group
with 500 connections.. (unrealistic, but for logic sakes)

does this make sense? or am I slowly going mad over this jmeter application?


Everything that has, did and will happen in your life, exists now, within
you (William Ottley)

Every universe is covered by seven layers — earth, water, fire, air, sky,
the total energy and false ego — each ten times greater than the previous
one. There are innumerable universes besides this one, and although they are
unlimitedly large, they move about like atoms in You. Therefore You are
called unlimited (Bhagavata Purana 6.16.37)

  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message