one FreeBSD server having NTP issues

So, this is super weird

I have 3 FreeBSD VMs on 3 separate ESXi servers that I use for storage with passthrough LSI 2008s

HW clocks are all set to local time, not UTC. TZ data are all set to PDT, which is where I live near Seattle.

They all connect to my Windows 2016 domain controller w/ samba and winbind, they all use krb5-116 and ntp

The krb5.conf and ntp.conf files are taken straight off the Samba wiki page for domain members and are working fine on two out of 3 of the VMs

Two of the VMs work great, one of them is having weird time issues. I wanted to think it was just a daylight savings time issue but it appears to be weirder than that -- it had been an hour ahead, but now it's a few minutes shy of that.

Working:
Code:
[root@fbsd111 /home/local]# date
Wed May 23 10:33:25 PDT 2018

vs problematic:
Code:
[root@fabby111 /home/local]# date
Wed May 23 11:30:58 PDT 2018


ntp.conf (same on all 3 VMs) - source https://wiki.samba.org/index.php/Time_Synchronisation
Code:
[root@fabby111 /home/local]# cat /etc/ntp.conf
server 127.127.1.0
fudge  127.127.1.0 stratum 10

server 192.168.1.2    iburst prefer
server 192.168.1.3    iburst

driftfile /var/lib/ntp/ntp.drift
logfile   /var/log/ntp

# Default restriction: Disallow everything
restrict default ignore

restrict 127.0.0.1

restrict 192.168.1.2    mask 255.255.255.255    nomodify notrap nopeer noquery
restrict 192.168.1.3    mask 255.255.255.255    nomodify notrap nopeer noquery

krb5.conf (same on all 3 VMs) - source: https://wiki.samba.org/index.php/Setting_up_Samba_as_a_Domain_Member#Configuring_Kerberos
Code:
[root@fabby111 /home/local]# cat /etc/krb5.conf
[libdefaults]
        default_realm = HAPPY.HUT
        dns_lookup_realm = false
        dns_lookup_kdc = true

ntp polling:
Code:
[root@fabby111 /home/local]# ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*LOCAL(0)        .LOCL.          10 l    4   64  377    0.000    0.000   0.000
 WINGATEWAY.happ .INIT.          16 u    - 1024    0    0.000    0.000   0.000
 1709dc02.happy. .INIT.          16 u    - 1024    0    0.000    0.000   0.000

Like I said, works fine for two out of 3 of the VMs - just started being an issue on the 3rd VM, which actually resides on the same ESXi host as my domain controller (which has no issues with time). It's an important VM for me because it's what I use for general file serving and physical computer backups - and I can't get the MFer to connect to the domain because of the clock skew:

Code:
[root@fabby111 /home/local]# kinit
administrator@HAPPY.HUT's Password:
kinit: krb5_get_init_creds: Clock skew too great

I just set the time using date command. I have no confidence that it will stay set correctly, though, and would like to work towards a more long-term solution.

Any ideas? Anyone?
 
Gah. I am now noticing, even though the ESXi host was using NTP for synchronizing time, the underlying manual date/time settings were set to the same time that my FreeBSD VM was defaulting to - must have set itself to the host's manual time clock instead of the NTP clock
he host was configured to use somehow.

I set the ESXi host clock manually by turning off NTP and then after setting it properly turned NTP back on (all on the ESXi host).

robotgirl_set_manual_time_also.PNG
t
 
Back
Top