9.16.3. KVM virtual machine timing management

Virtualization poses various challenges for virtual machine time keeping. Virtual machines which use the Time Stamp Counter (TSC) as a clock source may suffer timing issues as some CPUs do not have a constant Time Stamp Counter. Virtual machines running without accurate timekeeping can have serious affects on some networked applications as your virtual machine will run faster or slower than the actual time.
KVM works around this issue by providing virtual machines with a paravirtualized clock. The KVM pvclock provides a stable source of timing for KVM guests that support it.
Presently, only Red Hat Enterprise Linux 5.4 and higher virtual machines fully support the paravirtualized clock.
Virtual machines can have several problems caused by inaccurate clocks and counters:
  • Clocks can fall out of synchronization with the actual time which invalidates sessions and affects networks.
  • Virtual machines with slower clocks may have issues migrating.
These problems exist on other virtualization platforms and timing should always be tested.

Important

The Network Time Protocol (NTP) daemon should be running on the host and the virtual machines. Enable the ntpd service: 
# service ntpd start
Add the ntpd service to the default startup sequence: 
# chkconfig ntpd on
Using the ntpd service should minimize the affects of clock skew in all cases.
The NTP servers you are trying to use must be operational and accessible to your hosts and virtual machines.
Determining if your CPU has the constant Time Stamp Counter

Your CPU has a constant Time Stamp Counter if the constant_tsc flag is present. To determine if your CPU has the constant_tsc flag run the following command: 

$ cat /proc/cpuinfo | grep constant_tsc
If any output is given your CPU has the constant_tsc bit. If no output is given follow the instructions below.
Configuring hosts without a constant Time Stamp Counter

Systems without constant time stamp counters require additional configuration. Power management features interfere with accurate time keeping and must be disabled for virtual machines to accurately keep time with KVM.

Important

These instructions are for AMD revision F CPUs only.
If the CPU lacks the constant_tsc bit, disable all power management features (BZ#513138). Each system has several timers it uses to keep time. The TSC is not stable on the host, which is sometimes caused by cpufreq changes, deep C state, or migration to a host with a faster TSC. Deep C sleep states can stop the TSC. To prevent the kernel using deep C states append "processor.max_cstate=1" to the kernel boot options in the grub.conf file on the host: 
term Red Hat Enterprise Linux Server (2.6.18-159.el5)
        root (hd0,0)
	kernel /vmlinuz-2.6.18-159.el5 ro root=/dev/VolGroup00/LogVol00 rhgb quiet processor.max_cstate=1
Disable cpufreq (only necessary on hosts without the constant_tsc) by editing the /etc/sysconfig/cpuspeed configuration file and change the MIN_SPEED and MAX_SPEED variables to the highest frequency available. Valid limits can be found in the /sys/devices/system/cpu/cpu*/cpufreq/scaling_available_frequencies files.
Using the engine-config tool to receive alerts when hosts drift out of sync.

You can use the engine-config tool to configure alerts when your hosts drift out of sync.

There are 2 relevant parameters for time drift on hosts: EnableHostTimeDrift and HostTimeDriftInSec. EnableHostTimeDrift, with a default value of false, can be enabled to receive alert notifications of host time drift. The HostTimeDriftInSec parameter is used to set the maximum allowable drift before alerts start being sent.
Alerts are sent once per hour per host.
Using the paravirtualized clock with Red Hat Enterprise Linux virtual machines

For certain Red Hat Enterprise Linux virtual machines, additional kernel parameters are required. These parameters can be set by appending them to the end of the /kernel line in the /boot/grub/grub.conf file of the virtual machine.

Note

The process of configuring kernel parameters can be automated using the ktune package
The ktune package provides an interactive Bourne shell script, fix_clock_drift.sh. When run as the superuser, this script inspects various system parameters to determine if the virtual machine on which it is run is susceptible to clock drift under load. If so, it then creates a new grub.conf.kvm file in the /boot/grub/ directory. This file contains a kernel boot line with additional kernel parameters that allow the kernel to account for and prevent significant clock drift on the KVM virtual machine. After running fix_clock_drift.sh as the superuser, and once the script has created the grub.conf.kvm file, then the virtual machine's current grub.conf file should be backed up manually by the system administrator, the new grub.conf.kvm file should be manually inspected to ensure that it is identical to grub.conf with the exception of the additional boot line parameters, the grub.conf.kvm file should finally be renamed grub.conf, and the virtual machine should be rebooted.
The table below lists versions of Red Hat Enterprise Linux and the parameters required for virtual machines on systems without a constant Time Stamp Counter.
Red Hat Enterprise Linux Additional virtual machine kernel parameters
5.4 AMD64/Intel 64 with the paravirtualized clock Additional parameters are not required
5.4 AMD64/Intel 64 without the paravirtualized clock notsc lpj=n
5.4 x86 with the paravirtualized clock Additional parameters are not required
5.4 x86 without the paravirtualized clock clocksource=acpi_pm lpj=n
5.3 AMD64/Intel 64 notsc
5.3 x86 clocksource=acpi_pm
4.8 AMD64/Intel 64 notsc
4.8 x86 clock=pmtmr
3.9 AMD64/Intel 64 Additional parameters are not required
3.9 x86 Additional parameters are not required
Using the Real-Time Clock with Windows virtual machines

Windows uses the both the Real-Time Clock (RTC) and the Time Stamp Counter (TSC). For Windows virtual machines the Real-Time Clock can be used instead of the TSC for all time sources which resolves virtual machine timing issues.

To enable the Real-Time Clock for the PMTIMER clocksource (the PMTIMER usually uses the TSC) add the following line to the Windows boot settings. Windows boot settings are stored in the boot.ini file. Add the following line to the boot.ini file: 
/use pmtimer
For more information on Windows boot settings and the pmtimer option, refer to Available switch options for the Windows XP and the Windows Server 2003 Boot.ini files.
Report a bug