I have 2 HP ProLiant VMware 5.5 hosts and I have configured the time sync on them to use NTP. I have pointed the NTP to the n.north-america.pool.ntp.org servers. I have restarted the time service and waited 15 minutes but the time is still out by 6 minutes (currently anyway). The ntp.conf file looks like this:
restrict default kod nomodify notrap nopeer
restrict 127.0.0.1
server 0.north-america.pool.ntp.org
server 1.north-america.pool.ntp.org
server 2.north-america.pool.ntp.org
server 3.north-america.pool.ntp.org
driftfile /etc/ntp.drift
I have looked at the VMware NTP troubleshooting guide and the watch ntpq -q localhost command shows the following:
Every 2s: ntpq -p localhost
2016-05-20 17:39:28
remote refid st t when poll reach delay offset jitter
==============================================================================
+utcnist2.colora .NIST. 1 u 42 64 377 111.955 347166. 166.746
+nox.prolixium.c 200.98.196.212 2 u 42 64 377 56.777 347187. 163.313
triangle.kansas 128.252.19.1 2 u 114 64 332 105.921 347142. 155243.
*131.107.13.100 .ACTS. 1 u 40 64 377 103.846 347195. 165.405
The offset is always 347K out. I have also looked in syslog.log and the only messages I see are the following:
2016-05-20T16:36:58Z root: ntpd Stopping ntpd
2016-05-20T16:36:58Z watchdog-ntpd: Watchdog for ntpd is now 950158
2016-05-20T16:36:58Z watchdog-ntpd: Terminating watchdog process with PID 950158
2016-05-20T16:36:58Z watchdog-ntpd: [950158] Signal received: exiting the watchdog
2016-05-20T16:36:59Z ntpd[950186]: ntpd exiting on signal 1
2016-05-20T16:36:59Z root: ntpd Starting ntpd
2016-05-20T16:36:59Z watchdog-ntpd: [956340] Begin '/sbin/ntpd ++group=ntpd -g -n -c /etc/ntp.conf -f /etc/ntp.drift', min-uptime = 60, max-quick-failures = 1, max-total-failures = 100, bg_pid_file = ''
2016-05-20T16:36:59Z watchdog-ntpd: Executing '/sbin/ntpd ++group=host/vim/vmvisor/ntpd -g -n -c /etc/ntp.conf -f /etc/ntp.drift'
2016-05-20T16:36:59Z ntpd[956363]: ntpd 4.2.6p2@1.2194-o Tue Mar 3 09:11:43 UTC 2015 (1)
2016-05-20T16:36:59Z ntpd[956363]: proto: precision = 0.520 usec
2016-05-20T16:36:59Z ntpd[956363]: Listen and drop on 0 v4wildcard 0.0.0.0 UDP 123
2016-05-20T16:36:59Z ntpd[956363]: Listen normally on 1 lo0 127.0.0.1 UDP 123
2016-05-20T16:36:59Z ntpd[956363]: Listen normally on 2 vmk0 172.16.120.2 UDP 123
2016-05-20T16:36:59Z ntpd[956363]: Listen normally on 3 vmk1 172.16.180.23 UDP 123
2016-05-20T16:36:59Z ntpd[956363]: Listen normally on 4 vmk2 172.16.180.24 UDP 123
2016-05-20T16:42:01Z ntpd[956363]: 128.138.141.172 interface 172.16.180.23 -> 172.16.120.2
2016-05-20T16:42:01Z ntpd[956363]: 131.107.13.100 interface 172.16.180.23 -> 172.16.120.2
2016-05-20T16:42:01Z ntpd[956363]: 45.79.10.228 interface 172.16.180.23 -> 172.16.120.2
2016-05-20T16:42:01Z ntpd[956363]: 64.6.144.6 interface 172.16.180.23 -> 172.16.120.2
172.16.120.2 is my vmkernel port for vMotion, management, etc. 172.16.180.23 and 24 are vmkernel ports for iSCSI only (vMotion, management, etc. disabled).
There should be some messages in there similar to this "ntpd[263140]: synchronized to , stratum " but I don't see any.
Any idea of what I might be missing?
EDIT: In fact, if I manually set the time (using Google to get the current time) it will revert to 6 minutes slow after a few seconds.
EDIT 2: Well, the incorrect time was propagating to our domain controller and then out to our workstations. I disabled VMware Tools time sync to the domain controller, had it resync its time and then resync'd the time to my desktop so that everything was correct.
Now, I am not experiencing the above issue any longer. So, is it possible that the incorrect time on the domain controller VM was adjusting the time on the host? I didn't think it went both ways. Alternatively, would the vSphere Client on my machine update the time on the host?
