Forum Discussion

AndyHughes's avatar
AndyHughes
Regular Contributor
14 years ago

Another table log issue....

I've got a load test set up to out put the results at test step level in a table log. Which works.
There are 4 test steps 1. Data source 2. The request, 3. Properties and 4. Property transfer.

The load test is a ramp up with a limit set at 3000 requests. After about 2:30 the test ends and I save the results to a csv.

However the .csv shows an inconsistent number of test steps. For example there are 3779 rows which represent the test case as a whole and only 2530 representing the step which is the request. property trans = 2603, properties=2770 data source=3303
I could understand a little inconsistency, as by the end of the test the rate is around 40/sec. So maybe some don't finish etc. But we are talkin hundreds here.

Also, in the is file the 1st response for the request test step is recorded at 17:07:19:558

But in the Average table (which polls every half second) the 1st request which completes is recorded at 17:07:14:387, this is 5 seconds before the 1st response time in the other table. (and the requests take in the order of 20-30ms)

Any idea what's going on?
  • AndyHughes's avatar
    AndyHughes
    Regular Contributor
    Update....I've discovered that on a lot of occasions the row which represents the test case is duplicated maybe 2, 3, 4 or even 5 times with exactly the same time stamp, and because I also get the request out on the table log I can prove that the request is identical in each case. The request test step also repeats in some cases with exactly the same time stamp (although I can't prove these are the same as the request is linked to the test case rather than the test step, but I suspect that they are.)

    This leaves me in a mess with analysing the stats as I have to go through each file deleting duplicate rows in order to make sense of the data and make meaningful graphs etc.

    What this is doing to the other table I don't know? Do all of these duplicates get averaged into the results on the average table (thus skewing the results) or does it generate its results from elsewhere?
  • Andy,

    Please send me the project (or screenshots) and I'll have a look at this (issue created as [LUCO-2414]). I might need to know exactly how it's set up to faithfully reproduce it.

    Regards

    Henrik
    SmartBear Software
  • AndyHughes's avatar
    AndyHughes
    Regular Contributor
    Hi Henrick,

    It's the same set up as in my post "Assertion fails in loadui but not in soapui" apart from in this case it's a ramp generator rather than fixed rate.
    This is an example of a repeating row in the results file:



    Thanks