Help in understanding error messages

jupiter_sailormoon · May 6, 2020, 10:45am

I am running the following test

setUp(

OnePost.inject(rampUsers(noOfUsers) during (60 seconds)).protocols(httpProtocol),
postTwo.inject(nothingFor(150 second), rampUsers(noOfUsers) during (60 seconds)).protocols(httpProtocol),
OneGet.inject(nothingFor(200 second), rampUsers(noOfUsers) during (60 seconds)).protocols(httpProtocol),
getTwo.inject(nothingFor(200 second), rampUsers(noOfUsers) during (60 seconds)).protocols(httpProtocol)
)

In the above, I have the errors when I noOfUsers are 35000. I noticed also that the CPU was above 90%. Are the errors from the client or gatling?

Can anyone help me to understand the following error messages

slandelle · May 6, 2020, 11:29am

IO errors, most likely because you try to generate load from a machine that can’t handle it. Non beaffy laptop on Windows?

jupiter_sailormoon · May 6, 2020, 11:38am

Hi,
I am on an azure VM with4 vcpus, 16 GiB memory and 8 vcpus, 32 GiB memory to generate the load.

jupiter_sailormoon · May 6, 2020, 11:42am

Windows server 2019, VM on azure with 4 vcpus, 16 GiB memory and 8 vcpus, 32 GiB memory

slandelle · May 6, 2020, 12:29pm

Generally speaking, Java is as good on Windows as it is on Linux, in particular NIO
I hope you don’t have an anti-virus on this server…
We have tuning instructions, but for Linux (https://gatling.io/docs/current/general/operations/). Try finding similar instructions for Windows.
Linux machines are also available on Azure, if possible, I really recommend switching to Linux for Gatling.