NTP problems in nested ESXi setup

[EliSch]

Hi All,

I'm building nested ESXi setup for testing purposes:

My ESX 4.0.0, 208167 runs 4 VMs

Each one on the VMs has an ESXi 4.0.0, 208167 as guest OS

Everything works fine but the NTP clock synchronization for my ESXi hosts. As the result there is a significant clock drift that causes our application to fail.

The ESXIs are configured with correct NTP server and tcpdump on the ESXi shows that the NTP responses reach the ESXi host.

The VC colors the ESXi host's time in the "Configuration --> Time Configuration" menu in red.

The hosting ESX server is working fine with similar NTP settings.

Does anyone have any ideas on how to solve the issue?

Best Regards,

\Eli

[golddiggie]

Which NTP servers are you using??

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.

[EliSch]

We are using the 0.asia.pool.ntp.org

\Eli

[golddiggie]

Try adding to the time servers list. I'm using three NTP servers (all in the pool.ntp.org category).

1.asia.pool.ntp.org

2.asia.pool.ntp.org

3.asia.pool.ntp.org

I would also try a pool set that's a more grainular match to where you actually are... Such as the different countries listed here:

I've not used nested ESX/ESXi hosts before simply due to how many issues cropup that simply don't when you are on real hardware.

VMware VCP4

Consider awarding points for "helpful" and/or "correct" answers.

[EliSch]

Thank you for your answers.

It looks like the ESXi's NTP client adjusts the clock much slower then the ESX.

The time is red probably when there is "a lot to catch up".

Currently we have adjusted the clock manually, and it looks like the NTP does its fine tunning over the last 8 hours.

Best Regards,

\Eli

[EliSch]

OK, we have found the answer for this one:

adding the line:

timeTracker.apparentHz = 2376000000

in our ESXi VM vmx file helped to resolve the issue.

All credits goes to Harish and his team.