Dear,
I have a mixture of ESX 4 and ESXi5 hosts in my cluster.
They all use the same NTP servers (which are our 4 domain controllers)
But the time on the ESXi 5 servers is off by a 2-3 minutes.
I also really edit the time instead of just clicking refresh and seeing the red letters turn to black.
Anyone has an idea? Cause I think this is causing some issues, like performance graphs not showing up
Thanks
Erik
If the CPU of ESXi is over commited you will get into such issue. Once its back to normal the ESXI will start getting the correct time from the configured NTP server.
1. Make sure all the VM get time from DC, so that Vms dont expect this issue.
2. To avoid this issue in ESXI make sure you dont over commmit resources on it.
Also please make sure in VMware tools option time sync not should happened with ESXi Servers.
Well there are no machines running on the host yet.. since I wanted to fix my time sync problem first.
I remember in ESX4, when i entered an NTP server and started the service. The date & time corrected itself when clicking the refresh button.
In ESXi5 I can not get that to work. I really have to change the time myself manually to get the time from red to black.
Is that normal behavior?
1. Check if DNS is configured properly, Make sure there is not issue between ESXi5 and NTP server. if there is a firewall make sure the required port is open.
2. Try reinstalling the ESXi5 on the server and see if that fix.
It is not the normal behaviour, since you have told that without any VM there is a issue with time, there should be a CPU or mother board issue.
Run a diagonistic test on the esxi and log a call with hardware vendor with issue reporting this.
Check:
http://kb.vmware.com/kb/1005092 - general troubleshoot, only to be sure the NTP is ok.
http://kb.vmware.com/kb/1006090 - it should not be happening on 5, but check this out (not sure if the option is there)
If you unconfigure the NTP the problem have the same behavior? Can you confirm in BIOS that the time is running slow as well?
I have the same issue (see attach: red chars on NTP configuration). My network contains mixed ESXi 4.1.0 and 5.0.0 hosts.
A closer look reveal that (this evidence can be taken only on 5.0.0, since 4.10 does not have ntpdc) the dispersion value is extremely high
~ # ntpdc -ns
remote local st poll reach delay offset disp
=======================================================================
*192.168.64.10 192.168.63.102 1 64 377 0.00037 -0.649297 0.99852
.192.168.64.11 192.168.63.102 1 64 377 0.00037 -0.650600 0.99852
A quick comparison with (example below) working physical Linux machine show far better values
[root@luggage ~]# ntpdc -ns
remote local st poll reach delay offset disp
=======================================================================
.85.214.123.101 10.0.20.3 2 1024 377 0.06668 0.006177 0.13757
.95.129.55.109 10.0.20.3 3 1024 377 0.08525 0.004216 0.12390
*131.234.137.23 10.0.20.3 1 1024 377 0.10336 -0.016612 0.12398
Just some question:
Thanks in advance