ESXi fails to sync time with NTP server due to local system clock drift on HPE host
search cancel

ESXi fails to sync time with NTP server due to local system clock drift on HPE host

book

Article ID: 411423

calendar_today

Updated On:

Products

VMware vSphere ESXi

Issue/Introduction

  • Time of ESXi host is drifting heavily over a period of time with more and more larger offset from the output of ‘ntpq -p’:

         remote               refid          st   t     when  poll   reach   delay   offset  jitter
    ===============================================================================================
    *<ntp-server-ip-addr>    LOCAL(1)       10   u      59     64     377    0.183  -53.482  0.875

  • ESXi lost synchronization with NTP server finally in /var/run/log/vmkwarning.log

    vmkwarning: cpu103:2210480)WARNING: NTPClock: 1460: ntpd changed system clock to 1755591597.390714000, lost synchronization to upstream time servers

  • The local clock frequency error is beyond ntpd’s capabilities (500 ppm) that is lagging by times ranging over 1.44 seconds in /var/run/log/syslog.log

    sntp[211385]: 2025-08-18 06:20:04.735294 (+0000) +1.441665 +/- 0.999999 <ntp-server-ip-addr> s5 no-leap ######>>>>>> 1.441665 seconds

  • Linux OS installed on the host can successfully synchronize with the NTP server.

Environment

VMware vSphere ESXi 7.x
VMware vSphere ESXi 8.x

Cause

System time on different Operating Systems measures with different clock sources:

  • Linux uses the High Precision Event Timer (HPET)
  • Windows uses timers described in the Watchdog Action Table
  • ESXi bases on the TSC comparing with the ACPI PM Timer

The ACPI PM timer must run at a frequency close enough to the value specified by the ACPI standard that NTP could use it as a basis for the local clock. NTP cannot tolerate a frequency error greater than 500 ppm.

Resolution

It was determined that the XNC core clock frequency provided to BIOS which is used in setting the ACPI Power Management Timer frequency needed to be adjusted so that the ACPI Power Management Timer frequency was closer to the standard value.
This has been identified as an issue from HPE firmware. Engage HPE to get the latest firmware and BIOS.