The Monitoring Virtual Machines component is a function of the VMware High Availability (HA) tool in VMware VirtualCenter. For more information about this feature, see the
vSphere Availability Guide. Ensure that you have reviewed this information prior to proceeding with this test.
Environment Configuration
To configure your environment to use Monitoring virtual machines:
- Configure VMware HA on the cluster. For more information for your product version, see the Creating and Using VMware HA Clusters section in the vSphere Availability Guide.
Notes:
- vCenter Server 4.0 introduces an additional disk I/O status check, after virtual machine heartbeat fails to prevent accidental reboots.
- This ensures a wait of 120 seconds, of no disk I/O operation, before a restart.
- As a result, your virtual machine may not restart after completing the preceding steps.
To resolve this issue, disable the disk I/O check.
To disable the disk I/O check:
Set the das.iostatsInterval advanced vSphere HA attribute to 0.
- For more information on the das.iostatsInterval HA attribute, see the Customizing vSphere HA Behavior section in the vSphere Availability Guide for your product version.
-
Enable Monitoring Virtual Machines feature, and define a sensitivity setting. For more information, see the
Turn On Fault Tolerance for Virtual Machines in the vSphere Web Client section in the
vSphere Availability Guide for your product version.
-
-
Verify VM Restart Priority for the virtual machine does not have Disabled checkbox selected. For more information, see the
Selecting HA Options section in the
Resource Management Guide.
Test Procedure
This test simulates a loss of heartbeat signal from the virtual machine to replicate a monitoring failure. To create this loss of signal, you need to terminate the VMware Tools processes within the guest. When the heartbeat signal is lost, VMware HA monitors the signal for a predefined failure interval. After this failure interval has elapsed, VMware HA attempts to restart the virtual machine. For ESX hosts with hostd logging set to verbose, information is logged to the /var/log/vmware/hostd log files.
To disable the heartbeat, follow the procedure for the type of guest operating system installed on the virtual machine:
Windows Guest
You can stop the VMware Tools Heartbeat signal by stopping the VMware Tools service and VMwareUser.exe process within Windows.
Note: In vCenter Server5.0, the VMwareUser.exe process is replaced by vmtoolsd.exe.
In vCenter server 6.0, tools need to stop by following these steps:
- Check PID of vmtoolsd.exe by running this command:
"tasklist | findstr vmtoolsd"
- Kill all process of vmtoolsd.exe by running this command:
taskkill /PID/F
Linux Guest
Run this command to disable VMware Tools:
/etc/init.d/vmware-tools stop
Note: To test rest VM in 6.0
- Check PID of vmtoolsd by running this command:
#ps -alx |grep vmtoolsd
- Kill process of vmtoolsd by running this command:
#kill -9 PID
In addition, For RHEL 7.0 or later, If kill -9 command does not work, Please stop by running this command:
systemctl stop vmtoolsd.service