Interestingly, 'w32tm /monitor' return the same results and seem
to be connecting to the correct servers when working & non-working
clients are checked.
However checking the logon server (ie running 'set l') reveals
that the working clients are logging into the Primary DC and
the non-working clients apparently are loggin in to the backup DC.
I guess two questions:
1) Why isn't the logging into backup DC producing the sync'd time
2) How do the clients decide which machine to log in to?
On 6/27/2023 12:05 PM, Luis Peromarta via samba wrote:> I had a similar issue some time ago. Some clients would sync with DCs some
other didn?t.
>
> I recently had issues also with bookworm and NTP (bookwork installing
ntpsec). I decided to move all DCs to chrony, worked perfectly, and all member
servers to systemd-timesyncd with a minimal config from samba wiki that just
*works*
>
> Re. Windows clients, I have not seen may Local CMOS clocks lately after
these changes.
>
> This question may have been asked before, but how many DCs do you run, and
do all have the same config for time (ntp, chrony) ?
>
> What?s the output of??w32tm /monitor in the clients ? What?s the output of
w32tm /query /status ?
>
> On a good client (syncing with a DC) what is the logon server ? ( Run set l
)
> On a bad client (not syncing with a DC) what is the logon server ? Is there
a pattern ?
>
> Regards, LP
> On 26 Jun 2023 at 20:35 +0200, Marco Shmerykowsky PE <marco at
sce-engineers.com>, wrote:
>>
>> That's exactly what I thought I had done. The DC's point to
>> the netgate server. I was under the impression that the clients
>> would automatically time sync by logging into the domain.
>>
>> Some PC's/Clients are syncing. Others are connecting to "Local
>> CMOS Clock".