I don't understand why this problem occurs. There may be many reasons for this problem. I am not sure what causes this problem on my computer, so the method here is one of the methods that I occasionally try to get. It is not guaranteed to be effective.
In a word, the solution is to stop the ACPI service in the virtual OS.
Now let's talk about my environment and the process of discovery. My host environment is windows xp sp2, vmware 6.5 is installed, cent OS 5.3 is installed in vmware, xen 3.3.1 is installed in cent OS 5.3, then, install cent OS 5.3 on the xen domain0 and use putty in windows to operate the VM in vmware through ssh. The symptom is that virtual OS is often down, but vmware does not have any exception information. Open the task manager and find that the CPU usage of the vmware-vmx.exe process is 100%, and the memory usage is huge. Windows is also running slowly. Some people say they stopped the vmauthd service on google, but it didn't work for me. Later, we found that if there was no movement in the virtual OS for a period of time, this symptom would certainly appear. This is very similar to windows's standby mode, so it is likely that the acpi service in the virtual OS has played a role, enabling the operating system to enter the power-saving mode, at the same time, it also occupies a large amount of memory for recovery from the power-saving mode, thus affecting the windows system is also very slow. Therefore, go to the virtual OS (cent OS 5.3), run setup, and stop the acpid service. Then restart the virtual OS. Solve the problem: Wait for 40 minutes after the virtual ossclient is running, and then run the command again. The virtual ossclient still responds normally. The CPU usage of the vmware-vmx.exe process is normal, and the total CPU usage on the host is less than 10%.