Frequency of the NTP queries

wanradt · May 4, 2021 at 9:43 AM

Hi!

I use LibreElec (9.2.6 RPi2.arm) on Raspberry Pi 3b and after I included Pi-Hole in my network I noticed how many queries my LibreElec produces in my network. Specifically, there are on average about 20 NTP queries per minute. It seems to me a lot. I don't see the purpose of having so tight interval controlling the right time. Is this normal behavior? Can I somehow reduce the frequency of NTP queries? From my point of view once or twice in 24h should be enough.

Hope this is the right subforum here.

TIA,

Gunnar

**CvH** · May 4, 2021 at 10:40 AM Official Post

do you block the ntp queries or do they succeed?

if the system can't find a working ntp it is normal that it tries frequently

wanradt · May 4, 2021 at 12:28 PM

Quote from CvH

do you block the ntp queries or do they succeed?
if the system can't find a working ntp it is normal that it tries frequently

I don't block those queries, but I have no idea, how to control do they succeed or not

**chewitt** · May 4, 2021 at 1:51 PM Official Post

Code

May 04 13:25:57 N2PLUS connmand[524]: ntp: adjust (slew): +0.000828 sec
May 04 13:43:02 N2PLUS connmand[524]: ntp: adjust (slew): -0.001210 sec
May 04 14:00:06 N2PLUS connmand[524]: ntp: adjust (slew): +0.050135 sec
May 04 14:17:41 N2PLUS connmand[524]: ntp: adjust (slew): -0.078711 sec
May 04 14:34:45 N2PLUS connmand[524]: ntp: adjust (slew): +0.023562 sec
May 04 14:51:50 N2PLUS connmand[524]: ntp: adjust (slew): -0.000587 sec
May 04 15:08:55 N2PLUS connmand[524]: ntp: adjust (slew): +0.002952 sec
May 04 15:25:59 N2PLUS connmand[524]: ntp: adjust (slew): +0.000830 sec
May 04 15:43:03 N2PLUS connmand[524]: ntp: adjust (slew): +0.000999 sec
May 04 16:00:07 N2PLUS connmand[524]: ntp: adjust (slew): +0.001539 sec
May 04 16:17:11 N2PLUS connmand[524]: ntp: adjust (slew): +0.000139 sec
May 04 16:34:16 N2PLUS connmand[524]: ntp: adjust (slew): +0.000668 sec

Display More

^ running "journalctl | grep ntp" on an idle Amlogic device in the network here shows a check every 17 mins; more frequent than I would have expected but also nothing too extreme. It's not 20/minute though. Can you corroborate that frequency with journal entries?

**Klojum** · May 4, 2021 at 2:14 PM Official Post

I have Pi-Hole running as well. I'll have a check tonight on my RPi4, because there is some continious network stuff ongoing if I am to believe the UTP lights I'm not actively blocking them. But a NTP check every 3 seconds is a bit over the top...

wanradt · May 4, 2021 at 7:09 PM

Quote from chewitt

running "journalctl | grep ntp" on an idle Amlogic device

grepping ntp from journalctl gave me no output. Before the original posting, I looked into /var/log, but this directory is completely empty. Is there any logging mechanism I am not aware of?

To test do the ntp-queries succeed, I set the system clock an hour too early, but after running a few hours it remains wrong. So seems, that the problem may be in unsuccessful ntp-calls.

Then I checked LE Network settings and noticed that Timeserver#-settings were actually empty. I set first as pool.ntp.org and after restart, I got from journalctl also ntp-set logging:

Code

Apr 11 19:28:50 pildikast connmand[301]: ntp: adjust (jump): +65148582.224308 sec

This seems strange to me, but before setting up the timeserver-setting, it polled different nodes in pool.ntp.org but did not actually set the local clock. But now it does and does not pollute my Pi-Hole's log with not relevant queries.

Seems, we can close the case. Thank you all for the input!

Wbr,

Gunnar

**Klojum** · May 5, 2021 at 6:42 PM Official Post

For the record, on my RPi4 NTP is going off every 17m04s.

Code

May 05 18:21:12 rpi4c connmand[507]: ntp: adjust (slew): +0.007657 sec
May 05 18:38:16 rpi4c connmand[507]: ntp: adjust (slew): -0.013000 sec
May 05 18:55:20 rpi4c connmand[507]: ntp: adjust (slew): +0.004329 sec
May 05 19:12:24 rpi4c connmand[507]: ntp: adjust (slew): -0.008639 sec

**chewitt** · May 6, 2021 at 12:35 PM Official Post

It's odd. If you don't set any NTP servers via the LE settings add-on or connmanctl, connman will fall back to the embedded defaults which are some standard pool.ntp.org servers.

wanradt · May 6, 2021 at 1:05 PM

I agree, its strange. I can't reproduce the situation. Removing pool.ntp.org from settings does not cause now such flow of queries.

**chewitt** · May 6, 2021 at 5:49 PM Official Post

Computers

Frequency of the NTP queries

Similar Threads

Nightly images for A64, H3, H5, H6 and R40 boards

Rpi 4 B Audio Stutter

Libre Elec not booting (black screen) on Intel NUC