System clock drifts on ESXi 7.0 U3 or later after synchronization failure with NTP server.

Products

• VMware vSphere ESXi

Issue/Introduction

Symptoms -

• If there is a connectivity issue when attempting to reach an NTP server from an ESXi host, the system clock on the host may experience a significant drift.
• Running the vsish command vsish -e get /system/ntpclock/clockData will show a large negative value in the adjtime field.

NTP clock data {
  Clock state:0
  Clock status:0x00002001
  TAI offset:0
  Maximum error (usec):655420
  Estimated error (usec):349
  Poll shift:6
  Time offset (nsec):-143603
  Freq offset (nsec/sec):12888
  adjtime() (usec):-309237290312 <<<<<<
  tick adjustment (nsec/sec):-4987251
  Ref timer freq (cycles/sec):2399998000

Environment

• VMware vSphere ESXi 7.0
• VMware vSphere ESXi 8.0

Cause

This is a known issue in ESXi 7.0 U3 and later versions.

Resolution

The issue has been resolved in ESXi 7.0 Update 3o, ESXi 8.0 Update 1d, and ESXi 8.0 Update 2.

Workaround -

To resolve this issue, you can apply one of the following workarounds -

• Restore connectivity to the NTP server.
• Reset the ESXi clock state by running the following command - vsish -e set /system/ntpclock/clockData 1
• Confirm that the correct value is set in the adjtime() function by executing - vsish -e get /system/ntpclock/clockData
• The output should show - 

NTP clock data {
  Clock state:0
  Clock status:0x00000040
  TAI offset:0
  Maximum error (usec):503000
  Estimated error (usec):500000
  Poll shift:3
  Time offset (nsec):0
  Freq offset (nsec/sec):0
  adjtime() (usec):0 <<<<<
  tick adjustment (nsec/sec):0
  Ref timer freq (cycles/sec):2399998000
}

Additional Information

Impact/Risks -

• Failure to synchronize with the NTP server due to loss of connectivity leads to time drift on ESXi hosts.
• If the connectivity issue is temporary and subsequently resolved, time drift will be corrected eventually.