Some of our older Linux VPS may experience a fault after being live-migrated to a new host, where the system clock starts running much faster (or slower) than normal. The difference is so severe that even if an NTP daemon is running, it is unable to correct for the issue.


This fault occurs because some older Linux VPS have the "kvm-clock" source disabled, which is a special virtualized-clock that allows your VPS to read the time directly from the host hardware.


You can determine if you are NOT using kvm-clock by the following command:


$ cat /sys/devices/system/clocksource/clocksource0/current_clocksource
tsc


To correct this problem, go to the following page in mPanel:


Change your processor model to Intel KVM Processor v2  and then click Save and Apply.


After your VPS reboots, you can then confirm that kvmclock is in use:


$ cat /sys/devices/system/clocksource/clocksource0/current_clocksource

kvm-clock