PC12.53 Initialized TimeOut after LG OS upgrade

[Dennis]

Hi All,

Last week we installed a new server OS on 2 of our load generators.
After that I reinstalled the Performance Center Host 12.53 software again, including patch 04.
In ALM/Performance Center I removed the hosts and added them again after reinstalling the Performance Center Host software.

Now when I start a test from Performance Center, only aprox. 280 VU rampup to the RUN phase, but the other remaining 340VU all get timed out in the INIT phase and don't go to the RUN phase. I am not able to run  the test/scenario on the loadgenerators like I did before, when the loadgenerators were on a Windows 2008 R2(?) operating system.

What am I missing? Do I have to install something particular because it is a blade server (a lot of memory and CPU's)?
I think not, because the OS handles this and I see all the cores and memory in the Resource Monitor of Windows 2012 R2 on both load generators.
Do I have to do something particular to the Performance Center Host software due to this blade server?
There are no settings during the install of the Performance Center Host software.

I already tried a lot of things but haven't fixed it yet.
- checked the load generators system resources during the test execution
- checked the controller system resources during the test execution
- changed the temp folder on the generators
- changed the ramp-up of the test/scenario
- created a new test/scenario from scratch with the same settings as the existing test/scenario
- we did have a look at the anti-virus software, but there is not much disk IO during the the test execution

Can anybody give me some hints, new insights or help me with this issue?

Additional information:
Previous state of the load generators:
- 2 Physical blade servers, each server has 2 CPU's with 20 cores/hyper-threading(40CPU), 384GB ram, 64GB disk minimum
- Server OS Windows 2008 R2(?)
- Performance Center Host software 12.53 with patch 04
- running tests with 620VU TruClient FireFox

Current state of the load generators:
- 2 Physical blade servers, each server has 2 CPU's with 20 cores/hyper-threading(40CPU), 384GB ram, 64GB disk
- Server OS Windows 2012 R2
- Performance Center Host software 12.53 with patch 04
- fail to running tests with 620VU TruClient FireFox

Thanks in advance for thinking with me and/or helping me out.

Dennis

/Edit: some typos and we use TruClient FireFox

[Dennis]

Hi there,

We still have the same problem.

The next steps we took:
- excluded mdrv.exe, firefox.exe and other .exe file related to Performance Center/Loadrunner from the virusscanner via a Security Policy
- disabled Data Execution Prevention (DEP)
- changed the INIT time-out setting ( > 180 sec.)

We observe the following behavior when looking at the processes on the load generator:
- lrbridge.exe
   lr_host_balancer.exe
   - mdrv.exe
      firefox.exe
   - mdrv.exe
      firefox.exe
   - mdrv.exe
      firefox.exe
   - mdrv.exe
      firefox.exe
   - mdrv.exe
      firefox.exe
   - mdrv.exe
   - mdrv.exe
   - mdrv.exe
   - mdrv.exe
   - mdrv.exe
   - mdrv.exe
   - mdrv.exe

At a certain point, around 130VU on 1 load generator, the mdrv.exe process is started but not the firefox.exe process.
At this point the VU remains in the INIT phase on the controller and wil fail after the time-out period.

Is there a setting somewhere which prevent firefox.exe to start?
Is this a Performance Center/Loadrunner setting?
Is this an OS issue?

Dennis