jmeter-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From sebb <seb...@gmail.com>
Subject Re: Throughput calculation in distributed test is reporting txns/hours, not txns/seconds
Date Fri, 26 Feb 2010 02:37:21 GMT
On 26/02/2010, Michael Walker <michael.walker@amentra.com> wrote:
> I've been running tests using client/server mode, and may be running into
> bottlenecks on the test server, so I decided to add a second server and run
> a distributed test.
>
>  My original test typically runs at 160 tps. The two-node test appears to
> run at the same speed or faster. However, when I use the summary of
> aggregate report, the reported throughput is .1/hour.
>
>  I removed the second node from my jmeter-properties settings, and re-ran
> the test. Now, the same test reports a throughput of 80 tps, exactly half of
> the old throughput.
>
>  Does anyone have any idea what's going on here with these report
> calculations?

Unless the system clocks are reasonably well synchronised, combining
results can cause problems for throughput calculations, as the elapsed
time may be adversely affected.

Also, note that the same test is run on each server node, so running
with a single JMeter server instead of two will result in running half
the number of samples.

You may find it better to use two independent non-GUI runs and combine
the results later.

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

---------------------------------------------------------------------
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