Hi everyone,
I have been struggling with having the clock of Linux VM's (including CentOS 4.6 and 5.1) running very fast under VMWare, no matter what I have tried.
My platform: - AMD Turion X2 TL-60 - AMD 690 chipset with integrated Radeon x1250 (probably doesn't concern in this case) - Windows Vista Ultimate x64, all Windows Update patches loaded (no SP1 yet) - 4GB RAM - VMWare Workstation 6.0.2 - BIOS has no options to disable power management, etc.
On my host PC - C:\ProgramData\VMware\VMware Workstation\Config.ini, put in host.cpukHz = "2000000", host.noTSC = "TRUE", ptsc.noTSC = "TRUE" - AMD's dual core optimizer loaded
On CentOS 5.1 VM - use kernel options "noapic nolapic nosmp clocksource=acpi_pm" - use "divider=10" option in the regular kernel - try the vm version of the kernel in centos-plus with the same kernel options in first line
No matter what I try, the CentOS clocks are still fast - at a rate of almost 2 seconds for every 1 real second.
What other things I can do in order to fix this?
M. Chan
On Fri, Feb 22, 2008 at 12:12:59PM -0500, chanms wrote:
Hi everyone,
I have been struggling with having the clock of Linux VM's (including CentOS 4.6 and 5.1) running very fast under VMWare, no matter what I have tried.
My platform:
- AMD Turion X2 TL-60
- AMD 690 chipset with integrated Radeon x1250 (probably doesn't
concern in this case)
- Windows Vista Ultimate x64, all Windows Update patches loaded (no SP1 yet)
- 4GB RAM
- VMWare Workstation 6.0.2
- BIOS has no options to disable power management, etc.
On my host PC
- C:\ProgramData\VMware\VMware Workstation\Config.ini, put in
host.cpukHz = "2000000", host.noTSC = "TRUE", ptsc.noTSC = "TRUE"
- AMD's dual core optimizer loaded
On CentOS 5.1 VM
- use kernel options "noapic nolapic nosmp clocksource=acpi_pm"
- use "divider=10" option in the regular kernel
- try the vm version of the kernel in centos-plus with the same kernel
options in first line
No matter what I try, the CentOS clocks are still fast - at a rate of almost 2 seconds for every 1 real second.
What other things I can do in order to fix this?
I think the typical fix for the clock-too-fast problem is to boot your guest OS with the clock=pit option and then use the VMware Tools time synchronization option.
If the clock is running too slow is where you want to look into using one of the 100Hz kernels.
See: http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&exte...
Ray
On 2/22/08 9:20 AM, "Ray Van Dolson" rvandolson@esri.com wrote:
I think the typical fix for the clock-too-fast problem is to boot your guest OS with the clock=pit option and then use the VMware Tools time synchronization option.
If the clock is running too slow is where you want to look into using one of the 100Hz kernels.
See:
http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&exte... =1420
So the time sync option is meant to correct gross clock mismatches, but it's not intended to fix ongoing clock drift. If there's clock drift going on, it's probably another problem. Can you double-check that 'cat /sys/devices/system/cpu/cpu0/cpufreq/cpuinfo_max_freq' outputs the same number as your 'cpuKHZ' setting in your config.ini file? Have you also tried make sure that the CPU speed is locked to a constant speed? (See http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd... playKC&externalId=708 for help doing the latter - although it sounds like you may have already played around in our BIOS a little.)
Hope this helps, -- Elliot
Ray Van Dolson wrote: <snip>
I think the typical fix for the clock-too-fast problem is to boot your guest OS with the clock=pit option and then use the VMware Tools time synchronization option.
FWIW, I do this and the click still runs slightly too fast. I put this script in my cron.hourly folder and run the VMWare Tools Time Sync to fix any resulting slowness.
#!/bin/bash sleep 2 date --set='-1 second' >/dev/null 2>&1
I do no think the VMWare Tools handles "clock to fast" only "clock too slow". There must be a better way, but I do not know it.
If the clock is running too slow is where you want to look into using one of the 100Hz kernels.
I have found that the benefit of the VM kernel (aka 100Hz kernels) is that the load on the physical host is substantially reduced. I did not notice any timing issues/changes from this kernel.
Hi everyone,
I have been struggling with having the clock of Linux VM's (including CentOS 4.6 and 5.1) running very fast under VMWare, no matter what I have tried.
My platform: - AMD Turion X2 TL-60 - AMD 690 chipset with integrated Radeon x1250 (probably doesn't concern in this case) - Windows Vista Ultimate x64, all Windows Update patches loaded (no SP1 yet) - 4GB RAM - VMWare Workstation 6.0.2 - BIOS has no options to disable power management, etc.
On my host PC - C:\ProgramData\VMware\VMware Workstation\Config.ini, put in host.cpukHz = "2000000", host.noTSC = "TRUE", ptsc.noTSC = "TRUE" - AMD's dual core optimizer loaded
On CentOS 5.1 VM - use kernel options "noapic nolapic nosmp clocksource=acpi_pm" - use "divider=10" option in the regular kernel - try the vm version of the kernel in centos-plus with the same kernel options in first line
No matter what I try, the CentOS clocks are still fast - at a rate of almost 2 seconds for every 1 real second.
What other things I can do in order to fix this?
M. Chan
On 2/26/08 12:28 PM, "chanms" chanms@gmail.com wrote:
What other things I can do in order to fix this?
Hi,
There were a few replies to your post on Friday. Please follow up on them.
Best, -- Elliot
chanms wrote:
Hi everyone,
I have been struggling with having the clock of Linux VM's (including CentOS 4.6 and 5.1) running very fast under VMWare, no matter what I have tried.
My platform:
- AMD Turion X2 TL-60
- AMD 690 chipset with integrated Radeon x1250 (probably doesn't
concern in this case)
- Windows Vista Ultimate x64, all Windows Update patches
loaded (no SP1 yet)
- 4GB RAM
- VMWare Workstation 6.0.2
- BIOS has no options to disable power management, etc.
On my host PC
- C:\ProgramData\VMware\VMware Workstation\Config.ini, put in
host.cpukHz = "2000000", host.noTSC = "TRUE", ptsc.noTSC = "TRUE"
- AMD's dual core optimizer loaded
On CentOS 5.1 VM
- use kernel options "noapic nolapic nosmp clocksource=acpi_pm"
- use "divider=10" option in the regular kernel
- try the vm version of the kernel in centos-plus with the same kernel
options in first line
No matter what I try, the CentOS clocks are still fast - at a rate of almost 2 seconds for every 1 real second.
What other things I can do in order to fix this?
Did you try the 100Hz kernels in testing?
-Ross
______________________________________________________________________ This e-mail, and any attachments thereto, is intended only for use by the addressee(s) named herein and may contain legally privileged and/or confidential information. If you are not the intended recipient of this e-mail, you are hereby notified that any dissemination, distribution or copying of this e-mail, and any attachments thereto, is strictly prohibited. If you have received this e-mail in error, please immediately notify the sender and permanently delete the original and any copy or printout thereof.
-
chanms -
Elliot Lee -
John Thomas -
Ray Van Dolson -
Ross S. W. Walker