We are targetting a load test of 1200 users with 3 Load Generators (Of server configuration) and 1 Commander (Not Loading this machine) with each Load Generator configured to take 400 users.At around a load of 200 -250 users, we are getting the following 2 errors and the load on the load generator which gave these error comes down. So we see a drop in the active users. We tried restarting that task group involving the problematic Load Generator, one after the other it goes down giving the following error. We tried it in 1.4.0 and also in 1.4.2, similiar behaviour was observed in both the versions.
1. InjectorStatusImpl::GetInjectorTotalAndActiveVUs: Exception getting VU count from TG Executer 'xxx_xxx_xxx_xxx_00000680'; Could be TG stopped or completed
2. HTTPInjectorStatusImpl::HTTP_GetBothStatsSnapshot: Exception getting both snapshots from TG Executer 'xxx_xxx_xxx_xxx_00000680'; Could be TG stopped or completed
The Sampling Frequency and Task Monitoring Intervel are set to 30 Seconds. We tried reinstalling Opensta but the problem continues.
Has anybody encountered similiar problem?. Can anyone suggest a solution?.
Thanx in Advance.
It could be because of many reasons ,
The problem seems to be with the test machine, or it could be that your network is getting overloaded at high user load and STA is not able to communicate with test machines.
Typically OS has limitation on number of TCP/IP connections (2000 max), browser like IE are multithreaded applications opening multiple TCP/IP connections simultaneously, even this could account to the errors which you are getting at high userload.
One solution would be to use 6-8 test machines to effectively load you application, but keep a check on your network resources during the test.
A good workman is known by his tools.
If the only tool you have is a hammer, you tend to see every problem as a nail.