Score:1

Server

NTP client synchronizes with LOCAL instead of stratum 1

Moncloa

12/21/23, 9:25 AM

I have 2 NTP servers. One synchronized directly with a GPS clock (I will call this server 185), so st 0. Then I have a second one (Let's call it 186) that should synchronize with the 185 since it is st 1 but instead it synchronizes with LOCAL.

[root@185 ~]# ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*GENERIC(3)      .GPS.           0 l    6   64    7    0.000    5.902   2.454
 LOCAL(0)        .LOCL.          10 l    -   64    0    0.000    0.000   0.000

[root@186 ~]# ntpq -p
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
 GENERIC(3)      .GPS.           0 l    -   64    0    0.000    0.000   0.000
 192.168.61.185  .GPS.           1 u   52   64    1    0.238    0.036   0.000
*LOCAL(0)        .LOCL.          10 l   51   64    1    0.000    0.000   0.000

Configuration is quite straight forward since this is one of the first tests we do. In addition, I have the parameter tinker panic 0 set in both ntp.conf.

Do you have any idea why 186 does not choose 185?

Thanks in advance!

EDIT : To show that 186 can reach 185.

***EDIT : more info (date is 1 Jan because the GPS is simulated) ***

[root@ntp1 ~]# ntpdate -d 192.168.xx.185
 1 Jan 10:08:16 ntpdate[17348]: ntpdate [email protected] Fri Dec 18 13:55:10 UTC 2020 (1)
Looking for host 192.168.xx.185 and service ntp
host found : 192.168.xx.185
transmit(192.168.xx.185)
receive(192.168.xx.185)
transmit(192.168.xx.185)
receive(192.168.xx.185)
transmit(192.168.xx.185)
receive(192.168.xx.185)
transmit(192.168.xx.185)
receive(192.168.xx.185)
server 192.168.xx.185, port 123
stratum 1, precision -23, leap 00, trust 000
refid [GPS], delay 0.02576, dispersion 0.00000
transmitted 4, in filter 4
reference time:    c39e6ff9.17cf49c0  Thu, Jan  1 2004 10:07:53.093
originate timestamp: c39e7011.1621002b  Thu, Jan  1 2004 10:08:17.086
transmit timestamp:  c39e7011.16146846  Thu, Jan  1 2004 10:08:17.086
filter delay:  0.02586  0.02576  0.02576  0.02583
         0.00000  0.00000  0.00000  0.00000
filter offset: 0.000067 0.000094 0.000098 0.000065
         0.000000 0.000000 0.000000 0.000000
delay 0.02576, dispersion 0.00000
offset 0.000094

172

0 + 8

ntp

time-synchronization

ntpd

vidarlo

12/21/23, 9:36 AM

REACH is 0. That indicates that you can't reach that NTP server.

Moncloa

12/21/23, 10:12 AM

Actually it becomes 0 after a while, but after the restart reach isn't 0. `192.168.xx.185 LOCAL(0) 11 u 25 64 1 0.180 0.010 0.000`

vidarlo

12/21/23, 10:25 AM

[Edit] your post to include the output of `ntpdate -d 192.168.xx.185`. And stop redacting RFC1918-addresses. It just looks silly.

Moncloa

12/21/23, 10:31 AM

I do redact RFC1918-addresses because world is small, not for security issues... :)

vidarlo

12/21/23, 11:13 AM

Stop doing it anyway. It doesn't reveal anything about you.

Paul Gear

12/21/23, 11:04 PM

Please include your ntpd configuration as well - using `grep '^[^#]' /etc/ntp.conf` is recommended to make it compact.

Paul Gear

12/21/23, 11:04 PM

Also, as @vidarlo requested, please don't redact addresses. It just makes it harder for others to help you.

Paul Gear

12/21/23, 11:05 PM

And most importantly, please don't use the LOCAL driver. It has been deprecated for many years. Start with your distribution's default `/etc/ntp.conf` template (CentOS, Debian, & Ubuntu all have very sane defaults), and only change the things necessary for it to work in your environment.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: NTP client synchronizes with LOCAL instead of stratum 1

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.