29 August 2007

clock drift in Linux VMWare guest

Today I installed CentOS 5 as a VMWare guest (VMWare server, CentOS 4 host) and had a few problems. The first problem was that when it came time to partition the drive, CentOS didn't think I had any storage. A helpful post on the CentOS forums pointed out that I needed to select LSI (not BusLogic) for the SCSI controller.

And then I found that the clock drift was really bad, and NTP wasn't working for some reason.

http://kbase.redhat.com/faq/FAQ_43_9259.shtm suggested adding the following line to the .vmx file:
tools.syncTime = "TRUE"

This (by itself, anyway) didn't work for me.

http://www.djax.co.uk/kb/linux/vmware_clock_drift.html suggested appending the following items to the kernel command line (in lilo.conf or grub.conf):
nosmp noapic nolapic

That worked like a charm. That article also suggested appending 'clock=pit' if the guest clock runs fast (mine was running slow).

2 comments:

Joel Duckworth said...

tool.timesync = "TRUE" worked for me... why was this so hard to find... thanks

Anonymous said...

The links below may well offer a solution for you:


http://ubuntuforums.org/showthread.php?t=515539

http://blog.autoedification.com/2006/11/vmware-guest-clock-runs-fast.html


Cheers
kr8(at)bogus[dot]net