Vahan,
>I have configured my tests in the configuration pane of the OpenSTA
>commander so that there is a fixed delay time of 1 second between each
>iteration. Is it possible that VUs waiting for this delay are considered
>"inactive"??
Yes. Delays between script executions at the commander level would results
in a reduced number of VUs. If you export timers to a .csv file, how many
samples (response times) are exported? What percentage of them have the
number of active users less then the maximum number of VUs for the test?
Bernie Velivis
Performax Inc (www.iPerformax.com)
2 Colburn Lane
Hollis, NH 03049
603.860.7900
----- Original Message -----
From: "Vahan Harput" <vharput@...>
To: <opensta-users@...>
Sent: Monday, September 12, 2005 1:36 PM
Subject: [OpenSTA-users] Re: Re: What exactly is an "Active User"?
> >
>>
>>> Which VUs are considered "inactive" by OpenSTA?
>>
>> As far as I know, the number of active VUs is simply a count of the
>> number of threads that are running scripts. If you introduce users in
>> batches, then the number of VUs will vary over time according to the
>> number of users per batch, the batch run time, and the batch ramp up
>> interval. If a script exits with an error and no other script is
>> scheduled for that thread, then the number of active VUs will be
>> decremented.
>>
>> > Are those VUs executing a WAIT statement in a script considered
>> > "inactive"??
>>
>> No
>>
>> Regards,
>>
>> Bernie Velivis
>> Performax Inc (www.iPerformax.com)
>> 2 Colburn Lane
>> Hollis, NH 03049
>> 603.860.7900
>>
>
> During a test run, after the ramp up time, where the no. of active VUs
> reaches the maximum, this number will temporarily decrease and after a
> while, increase to the maximum again. This happens repeatedly during the
> whole test run. Interestingly, the error log does not contain any errors.
> The audit log will sometimes (but not always!) contain the message "E*
> HTTPRESPONSE (HTML parser): unresolved variable for connection (X)" if
> the value of a variable could not be extracted from the response (happens
> usually if the stressed application throws an exception). However, I don't
> think that an unresolved variable will cause a VU to be "inactive" since I
> assume that the VU will continue executing the script even if a variable
> could not be resolved.
>
> I have configured my tests in the configuration pane of the OpenSTA
> commander so that there is a fixed delay time of 1 second between each
> iteration. Is it possible that VUs waiting for this delay are considered
> "inactive"??
>
> Regards,
> Vahan Harput
>
>
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle
> Practices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> --
> OpenSTA-users mailing list OpenSTA-users@...
> Subscribe/Unsubscribe/Options:
> http://lists.sf.net/lists/listinfo/opensta-users
> Posting Guidelines:
> http://portal.opensta.org/faq.php?topic=UserMailingList
|