LRE Host machines, used as Controllers, are becoming non-operational frequently

Category:

Load/Performance Testing

Question ID: 109548

For the last one we are facing issue with controller machines and going non-operational very frequently, when we start running a load test. The same machines worked without any issue for the last five months. Installed performance center host on all these machines and being used as controller and data processor.

All the required services are running:
§ DataCollectionAgent
§ Micro Focus Network Virtualization (if installed)
§ RemoteManagement Agent Service
§ Performance Center Agent Service
§ Performance Center Load Testing Service
§ Performance Center Analytics
§ Performance Center Data Service
§ Secondary Logon

The Performance System User (Default=IUSR_METRO) exists in the Administrators group.

The Performance System User (Default=IUSR_METRO) and the Administrators are both defined in the “Users Rights Assignment”, in the Local Policy Editor. In the following policies:
§ Access this computer from the network
§ Allow log on locally
§ Create global objects
§ Impersonate a client after authentication
§ Log on as a batch job
§ Log on as a service
§ Replace a process level token

From the PC Host’s LTOPSvc log, I found the following error message:

2020-05-01 05:52:19,440 [6116] [23] ERROR HP.PC.Common.Utilities.ProcessHelper – Failed to kill process Wlrun, id 9300. [Context: 9f6c98fe-623d-440c-ae06-7e42825635e0] [FullContext: QCServer:http://ALMSERVERNAME:8080/qcbin UserName:_pc_system_ Domain:DEFAULT Project:LAB_PROJECT]
System.ComponentModel.Win32Exception (0x80004005): Access is denied
at System.Diagnostics.Process.Kill()
at HP.PC.Common.Utilities.ProcessHelper.KillProcess(Process process)
2020-05-01 05:52:19,440 [6116] [23] ERROR HP.PC.LTOP.Core.MaintenanceTasks.ResourceRecoveryOperation – Failed to launch controller. Error: Failed to kill existing controller [Context: 9f6c98fe-623d-440c-ae06-7e42825635e0] [FullContext: QCServer:http://ALMSERVERNAME:8080/qcbin UserName:_pc_system_ Domain:DEFAULT Project:LAB_PROJECT]
HP.PC.LT.Model.Exceptions.LaunchCntrlException: Failed to kill existing controller

Any ideas of what can be the cause?

Marked as spam

Posted by testrabbit (Questions: 144, Answers: 1)

Asked on May 18, 2020 10:31 am

488 views

Answers (1)

▲

▼

✔

Private answer

This error message indicates that security, on the PC Host machine is blocking the Wlrun process, by denying the System User (IUSR_METRO) permissions in the "Access is denied".

I researched the "Failed to kill process Wlrun" and "Error: Failed to kill existing controller" error messages on the Micro Focus Self-Solve Knowledge database, and found the following KM article KM03158599:

https://softwaresupport.softwaregrp.com/document/-/facetsearch/document/KM03158599

If you have FlexNet Inventory Agent 14.10.22 installed on your LRE Host machines, this is the root cause of your PC Host becoming non-operational.

A Google search also confirmed that "FlexNet Inventory Agent" is the agent for the Flexera Inventory Manager security service.

Solution:

1. Stop and disable FlexNet Inventory Agent
2. Once the service is stopped kill wlrun.exe in the Task Manager.
3. Reconfigure the PC Host and it will switch to an operational state.
4. Exclude ALL Performance Center host from the Flexera Inventory Manager security service scan( FlexNet Inventory Agent) to prevent the issue from happening in the future.

Please contact Flexera support or your Windows/Security team for instructions on how it can be excluded.

Marked as spam