jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Michael Stover <mstov...@apache.org>
Subject Re: Beginner's questions
Date Thu, 04 Nov 2004 23:08:54 GMT
On Thu, 2004-11-04 at 06:00, Jens Peter Christensen wrote:
> 
[snip]
> 1) The recording creates many HTTP Requests for the same GIF files. I
> suppose these are conditional gets that checks if the file has changed
> since last get, but I'm unsure if this is retained in the requests. Will I
> need to delete all redundant HTTP Request samples? If so, is there an easy
> way to do this? I want to keep the initial get, since this will be included
> in a real-world scenario.

In the Proxy Recording Server, you can make regexes to filter out file
types you don't wish to record.  You can also choose to include only
certain file types.  I usually prefer not to record the static resources
on a dynamic web app, as they don't typically account for much of the
load.  In any case, they get recorded as often as your browser requests
them, so if your browser isn't caching them, either caching is turned
off, or there's a problem.

> 
> 2) The requests all include server name and port number, but I would like
> to be able to use the test against different server (with the same
> application) to compare performance. I understand that the "HTTP Request
> Defaults" config element only applies to samples where the servername and
> portnumber entries are empty. How do I remove these entries from all
> samples easiest?

Create User Defined Variables in the Test Plan element for "server" and
give it the value of the host your recording against.  JMeter will
replace all instances of the host name with ${server}.  You could do
this for port as well, but you may find that replacing every occurrence
of "80" in your requests is not what you want to do (any parameters
whose value includes "80" would have a ${port} embedded in them).

For port, if your recording against port 80, this field should be left
blank by JMeter, and you can set it globally using an HTTP Request
Defaults component.

> 
> 3) How would I simulate a realworld user scenario with suitable pauses?
> Would this be something like a thread group with some timer to simulate
> waittime between ckicks, and then have all requests bundled into a number
> of simple controllers, each with the requests that are fired by a
> user-click and a timer with value 0?

I assume you're referring to the embedded resources that get recorded as
separate requests.  No good way to deal with that.  What I have done in
the past is attach a gaussian random timer to the initial page request
object, and no timers on the others.  Since a timer fires before it's
requests, this will delay the initial page request.  If you attach a
timer at the threadgroup level, it will delay all requests in the
threadgroup, which may not be what you want.

-Mike
> 
> thanks for any thoughts,
> Jens Peter Christensen, IRD, MPL
> Intentia A/S
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: jmeter-user-help@jakarta.apache.org
-- 
Michael Stover <mstover1@apache.org>
Apache Software Foundation


---------------------------------------------------------------------
To unsubscribe, e-mail: jmeter-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: jmeter-user-help@jakarta.apache.org


Mime
View raw message