-1

We have a cluster where exact time synchronization is very important. We use Ubuntu 20.04 servers and and chrony. Unfortunately there are always problems with time synchronization. No other known time synchronization service is running. I can also provide the Chrony logs It would be very nice if someone can help me.

host1-mstr01 ntpq environment: line 7: ntpq: command not found

ntpstat environment: line 10: ntpstat: command not found

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:50 CET Universal time: Tue 2020-12-22 08:43:50 UTC RTC time: Tue 2020-12-22 08:43:50 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:43:41 2020 System time : 0.000016022 seconds slow of NTP time Last offset : -0.000016416 seconds RMS offset : 0.000014285 seconds Frequency : 34.582 ppm slow Residual freq : +0.002 ppm Skew : 0.042 ppm Root delay : 0.004369508 seconds Root dispersion : 0.016251557 seconds Update interval : 514.3 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 8 377 9 -7039ns[ -23us] +/-
34ms

host2-mstr01 ntpq environment: line 7: ntpq: command not found environment: line 10: ntpstat: command not found

ntpstat

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead) since Tue 2020-09-29 17:46:17 CEST; 2 months 22 days ago Main PID: 12619 (code=exited, status=0/SUCCESS) Status: "Shutting down..."

Aug 11 13:31:33 host2-mstr01 systemd[1]: Started Network Time Synchronization. Aug 11 13:31:33 host2-mstr01 systemd-timesyncd[12619]: Initial synchronization to time server 10.0.16.254:123 (host1-NTP01). Sep 17 10:44:52 host2-mstr01 systemd-timesyncd[12619]: Timed out waiting for reply from 10.0.16.254:123 (host1-NTP01). Sep 17 13:00:51 host2-mstr01 systemd-timesyncd[12619]: Initial synchronization to time server 10.0.16.254:123 (host1-NTP01). Sep 21 04:03:23 host2-mstr01 systemd-timesyncd[12619]: Initial synchronization to time server 10.0.16.254:123 (host1-NTP01). Sep 26 10:58:27 host2-mstr01 systemd-timesyncd[12619]: Initial synchronization to time server 10.0.16.254:123 (host1-NTP01). Sep 27 02:49:01 host2-mstr01 systemd-timesyncd[12619]: Initial synchronization to time server 10.0.16.254:123 (host1-NTP01). Sep 29 17:46:17 host2-mstr01 systemd[1]: Stopping Network Time Synchronization... Sep 29 17:46:17 host2-mstr01 systemd[1]: systemd-timesyncd.service: Succeeded. Sep 29 17:46:17 host2-mstr01 systemd[1]: Stopped Network Time Synchronization.

timedatectl Local time: Tue 2020-12-22 09:43:51 CET Universal time: Tue 2020-12-22 08:43:51 UTC RTC time: Tue 2020-12-22 08:43:51 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:33:06 2020 System time : 0.000032089 seconds fast of NTP time Last offset : +0.000019423 seconds RMS offset : 0.000798126 seconds Frequency : 2.984 ppm slow Residual freq : +0.000 ppm Skew : 0.036 ppm Root delay : 0.018379021 seconds Root dispersion : 0.010704475 seconds Update interval : 1040.9 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 10 377 644 +78us[ +97us] +/-
32ms environment: line 7: ntpq: command not found

host1-node01 ntpq

ntpstat environment: line 10: ntpstat: command not found

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:52 CET Universal time: Tue 2020-12-22 08:43:52 UTC RTC time: Tue 2020-12-22 08:43:52 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:42:25 2020 System time : 0.000003835 seconds fast of NTP time Last offset : +0.000009851 seconds RMS offset : 0.000719458 seconds Frequency : 34.179 ppm slow Residual freq : +0.014 ppm Skew : 0.088 ppm Root delay : 0.004465697 seconds Root dispersion : 0.021002445 seconds Update interval : 128.5 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 7 377 86 +14us[ +24us] +/-
33ms environment: line 7: ntpq: command not found

host1-node02 ntpq environment: line 10: ntpstat: command not found

ntpstat

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:53 CET Universal time: Tue 2020-12-22 08:43:53 UTC RTC time: Tue 2020-12-22 08:43:53 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:42:28 2020 System time : 0.000008563 seconds slow of NTP time Last offset : -0.000016512 seconds RMS offset : 0.002114891 seconds Frequency : 32.695 ppm slow Residual freq : -0.029 ppm Skew : 0.250 ppm Root delay : 0.004427484 seconds Root dispersion : 0.021170758 seconds Update interval : 130.4 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 7 377 84 -16us[ -32us] +/-
33ms environment: line 7: ntpq: command not found

host1-node03 ntpq

ntpstat environment: line 10: ntpstat: command not found

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:55 CET Universal time: Tue 2020-12-22 08:43:55 UTC RTC time: Tue 2020-12-22 08:43:55 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:42:16 2020 System time : 0.000002321 seconds fast of NTP time Last offset : -0.000001159 seconds RMS offset : 0.000047930 seconds Frequency : 34.080 ppm slow Residual freq : +0.001 ppm Skew : 0.060 ppm Root delay : 0.006130085 seconds Root dispersion : 0.017225547 seconds Update interval : 519.0 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 9 377 99 -3398ns[-4557ns] +/-
33ms

host2-node01 ntpq environment: line 7: ntpq: command not found

ntpstat environment: line 10: ntpstat: command not found

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:56 CET Universal time: Tue 2020-12-22 08:43:56 UTC RTC time: Tue 2020-12-22 08:43:56 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:29:07 2020 System time : 0.000026769 seconds fast of NTP time Last offset : +0.000061154 seconds RMS offset : 0.000095979 seconds Frequency : 1.427 ppm slow Residual freq : +0.001 ppm Skew : 0.021 ppm Root delay : 0.017465375 seconds Root dispersion : 0.011619666 seconds Update interval : 1044.3 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 10 377 889 +380us[ +441us] +/-
27ms environment: line 7: ntpq: command not found

host2-node02 ntpq environment: line 10: ntpstat: command not found

ntpstat

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:57 CET Universal time: Tue 2020-12-22 08:43:57 UTC RTC time: Tue 2020-12-22 08:43:57 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:31:51 2020 System time : 0.000060706 seconds fast of NTP time Last offset : +0.000152064 seconds RMS offset : 0.000274430 seconds Frequency : 1.858 ppm slow Residual freq : +0.005 ppm Skew : 0.116 ppm Root delay : 0.019006269 seconds Root dispersion : 0.018152438 seconds Update interval : 1030.5 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 10 377 726 +297us[ +449us] +/-
30ms environment: line 7: ntpq: command not found

host2-node03 ntpq

ntpstat environment: line 10: ntpstat: command not found

● systemd-timesyncd.service Loaded: masked (Reason: Unit systemd-timesyncd.service is masked.) Active: inactive (dead)

timedatectl Local time: Tue 2020-12-22 09:43:59 CET Universal time: Tue 2020-12-22 08:43:59 UTC RTC time: Tue 2020-12-22 08:43:59 Time zone: Europe/Berlin (CET, +0100) System clock synchronized: yes NTP service: active RTC in local TZ: no

chronyc tracking Reference ID : 0A0010FE (host1-timeserver) Stratum : 3 Ref time (UTC) : Tue Dec 22 08:38:06 2020 System time : 0.000103265 seconds fast of NTP time Last offset : +0.000101621 seconds RMS offset : 0.000754228 seconds Frequency : 2.710 ppm slow Residual freq : +0.002 ppm Skew : 0.065 ppm Root delay : 0.018792789 seconds Root dispersion : 0.015140322 seconds Update interval : 1035.0 seconds Leap status : Normal

chronyc sources 210 Number of sources = 1 MS Name/IP address
Stratum Poll Reach LastRx Last sample =============================================================================== ^* host1-timeserver 2 10 377 353 +342us[ +443us] +/-
36ms

tom
  • 29
  • 1
  • 1
  • 2
  • You say there are always problems, but you don't say which problems. Can you elaborate? Also please format your code so that it becomes readable, and point out which parts of your code dump are relevant (or just cut out the irrelevant parts). – berndbausch Dec 23 '20 at 04:25
  • The problem is that we have synchronisations problem in the cluster.In the cluster chrony is used . The timeserver is ntp.service. Above I postet the output from every host in the cluster form this commands: timedatectl, chronyc tracking, chronyc sources and others. I can not see a mistake. But still we have time synchronisations problems (clock skew). Do have any any Idea where the problem is? Shall I provide the chrony logs? – tom Dec 23 '20 at 11:48

1 Answers1

0

I set up an extra time server for the cluster. The clients in the cluster now synchronize with the time server every second. The problem has not occurred since then. I guess the problem is solved.

server clustertimeserver minpoll 0 maxpoll 0
(using one second polling interval)
tom
  • 29
  • 1
  • 1
  • 2