Unanswered question

Runtime errors freezed for long running test

During long running test like Soak, we observed that, the runtime errors screen is freezed and couldn't get the latest errors. We had to stop the test and re-start again.

It occurs on v7.0.0. Is there any known bug on this version?

Jabastin P.
Jabastin P.

Jabastin P.

Level
0
37 / 100
points

Answers

How many errors did you see? I believe there is a buffer limit of 20,000 errors. This can be increase but it might be you just filled this up? just the fact that you said it was a long running soak test

Jabastin P.
Jabastin P.

Jabastin P.

Level
0
37 / 100
points

I can see around 30k errors in my soak test so far. Yes, all the errors were from our 4-5 days long running test. How can we increase the limit please?

Ideally you should look to deal with these errors first, are you using try/catch in your tests as most of these might be fallout from earlier errors and not genuine issues. I will find the key later as i am on site at the moment

Jabastin P.
Jabastin P.

Jabastin P.

Level
0
37 / 100
points

Yes, the errors were expected from my test. As we are trying to see the stability for 5 days continuous long running test with minimal load on my application. During the test, we would want to see what errors are occurring frequently / occasionally based on some quick fix from my application developers.

In the controller.properties file you will see in the Results section...

max.errors.store.count=20000

(It is the number of error to store on disk on Controller side. When this total number is received, Controller sends to each LG an order to limit sending error entries)

The new setting will only take effect after Neoload is restarted