jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Deepak Shetty <shet...@gmail.com>
Subject Re: Test plan for (970 page requests every 5 min)
Date Sat, 11 Sep 2010 01:08:06 GMT
>Ok. What is the sign that my server can't handle the burst?
a. Socket errors on server. Unix errors like too many open sockets /
handles.
b. Status code 503 being returned as HTTP response code (or 500 or whatever
its using)
c. CPU on server being pegged at 100% (may even be out of memory exceptions
on the server)
d. Extremely slow response times

>Let's say I have 1000 samples in a thread group (no think times etc). Are
all 1000 sent in 1 ms.
A thread groups  also has a number of  threads right. Each thread will
request a sample in the order in the test plan , get the response , finish
processing it (i.e. running post processors assertions , notifying listener
etc) and then request the next sample. i.e. each thread sends requests
sequentially. The time it takes to send 1000 samples depends on the time the
server takes to process the request and return a response back to JMeter.

However since you normally have multiple threads , each of which is doing
the same thing independently (depends on test plan and controllers) , the
server will see requests in parallel (with a concurrency maximum = number of
threads). However since the server will not always respond exactly the same
, its possible that Thread1 is on request 100 whereas thread2 is on request
110 and thread3 is on request 90.

regards
deepak

On Fri, Sep 10, 2010 at 5:31 PM, Prostak <strodion@gmail.com> wrote:

>
> >You give a rampup because you dont know whether your server can handle the
> >burst - you can give a rampup even if the server could handle the burst
> (so
> >give it irrespective).
>
> Ok. What is the sign that my server can't handle the burst? I don't even
> know what to look for... How do I tell?
>
>
> Can you also please answer this question: Let's say I have 1000 samples in
> a
> thread group (no think times etc). Are all 1000 sent in 1 ms? Or are they
> sent one by one over a period of time? I am trying to understand Jmeter's
> logic.
>
>
> --
> View this message in context:
> http://jmeter.512774.n5.nabble.com/Test-plan-for-970-page-requests-every-5-min-tp2826174p2835740.html
> Sent from the JMeter - User mailing list archive at Nabble.com.
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
>
>

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