ntpd both running and not running?

[danb35]

I could, of course, try to reboot the system--but that sounds like an awfully Windows-y troubleshooting step.

Well, not having anything else to try, I did this.  No change; ntpdate process continue to spawn, ntpd won't start, and /var/log/ntpd.log isn't updating.  I do note that the reported time is off by -4 hours, which coincidentally enough is my offset from UTC.

Edit: I just upgraded to 20.7, so I'm not sure if I should continue the thread here.  But here's what the log looked like after the reboot:

Code Select Expand


Jul 31 09:16:51 opnsense ntpdate[12533]: Can't find host 0.pool.ntp.org: Name does not resolve (8)
Jul 31 09:16:54 opnsense ntpdate[12533]: Can't find host 2.pool.ntp.org: Name does not resolve (8)
Jul 31 09:16:56 opnsense ntpdate[12533]: no servers can be used, exiting
Jul 31 09:17:25 opnsense ntpdate[94626]: Can't find host 0.pool.ntp.org: Name does not resolve (8)
Jul 31 13:17:54 opnsense ntpdate[94626]: step time server 192.241.146.233 offset +14409.531949 sec
Jul 31 13:17:54 opnsense ntp[16779]: Successfully synced time after 2 attempts.
Jul 31 13:17:54 opnsense ntp[78028]: Starting NTP Daemon.
Jul 31 13:17:54 opnsense ntpd[48610]: ntpd 4.2.8p15@1.3728-o Tue Jul 28 02:25:36 UTC 2020 (1): Starting
Jul 31 13:17:54 opnsense ntpd[48610]: Command line: /usr/local/sbin/ntpd -g -c /var/etc/ntpd.conf -p /var/run/ntpd.pid
Jul 31 13:17:54 opnsense ntpd[48610]: ----------------------------------------------------
Jul 31 13:17:54 opnsense ntpd[48610]: ntp-4 is maintained by Network Time Foundation,
Jul 31 13:17:54 opnsense ntpd[48610]: Inc. (NTF), a non-profit 501(c)(3) public-benefit
Jul 31 13:17:54 opnsense ntpd[48610]: corporation.  Support and training for ntp-4 are
Jul 31 13:17:54 opnsense ntpd[48610]: available at https://www.nwtime.org/support
Jul 31 13:17:54 opnsense ntpd[48610]: ----------------------------------------------------
Jul 31 13:17:54 opnsense ntpd[55576]: proto: precision = 0.288 usec (-22)
Jul 31 13:17:54 opnsense ntpd[55576]: basedate set to 2020-07-16
Jul 31 13:17:54 opnsense ntpd[55576]: gps base set to 2020-07-19 (week 2115)
Jul 31 13:17:54 opnsense ntpd[55576]: restrict: 'monitor' cannot be disabled while 'limited' is enabled
Jul 31 13:17:54 opnsense ntpd[55576]: Listen and drop on 0 v6wildcard [::]:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen and drop on 1 v4wildcard 0.0.0.0:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen normally on 2 igb2 [fe80::208:a2ff:fe0a:d506%3]:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen normally on 3 igb2 192.168.1.1:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen normally on 4 lo0 [::1]:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen normally on 5 lo0 127.0.0.1:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen normally on 6 igb2_vlan107 [fe80::208:a2ff:fe0a:d506%9]:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listen normally on 7 igb2_vlan107 192.168.107.1:123
Jul 31 13:17:54 opnsense ntpd[55576]: Listening on routing socket on fd #28 for interface updates
Jul 31 13:18:53 opnsense ntpd[55576]: ntpd exiting on signal 15 (Terminated)
Jul 31 13:18:53 opnsense ntpd[55576]: 162.159.200.1 local addr 192.168.1.1 -> <null>
Jul 31 13:18:53 opnsense ntpd[55576]: 23.131.160.7 local addr 192.168.1.1 -> <null>
Jul 31 13:18:53 opnsense ntpd[55576]: 2600:3c01::f03c:92ff:fe12:add4 local addr fe80::208:a2ff:fe0a:d506%3 -> <null>
Jul 31 13:18:53 opnsense ntpd[55576]: 90.187.99.165 local addr 192.168.1.1 -> <null>


It looks to me like ntpdate_sync_once.sh completed successfully, ntpd started, and about a minute later, something killed it.  And I'm back to ntpdate starting over and over, and nothing at all being written to ntpd.log.

[danb35]

Any other ideas?  Or since this problem is still present in 20.7, should I post this issue in that forum?

[mimugmail]

How about new chrony plugin (devel)? Go to CLI and type "pkg install os-chrony-devel".
The just disable ntp and do the rest in chrony .. should be more solid than ntpd and I need some beta testers ;)

[danb35]

Since this looks to me like a bug, reported here:
https://github.com/opnsense/core/issues/4245

[danb35]

How about new chrony plugin (devel)?

I'll consider it, but I'd still like to figure out why this one isn't working.  What does Chrony do differently that might make it a better solution?

[danb35]

So apparently the official answer (https://twitter.com/opnsense/status/1293624624606072834 - it's on Twitter, so it must be official, right?) is "we don't know what's going on, we don't know how to track it down, and we can't be bothered to try to figure it out."  Disappointing.

For some updates--the problem is present in 20.1.9 on a Netgate RCC-VE-2440.  It's still there when I upgrade that installation to 20.7.  It's also present on a clean install of 20.7 on a Protectli FW4B, with the configuration file imported from the previous system.  So maybe it's something that's only affecting me, but it's doing it across two different releases of the software and on two different systems.

So, since this seems to be at a dead end, on to Chrony.  It installs just fine.  It starts up and appears to run without errors, which puts it ahead of ntpd.  I'm noticing what seems to be one bug, though--there's a field in the GUI for "Allowed Networks", but it only lets me enter one network--when I enter a second one, I get a validation error.

[franco]

> we don't know what's going on, we don't know how to track it down, and we can't be bothered to try to figure it out.

To be frank, it's not what the Twitter response actually says.

Looking at https://blog.fosketts.net/2015/09/21/the-ideal-pfsense-platform-netgate-rcc-ve-2440/ it would seem that "Rangeley dual-core Atom 1.7 GHz CPU" is pretty thin and the Twitter response already states that both Ntpd and OpenVPN issues may be related to a sluggish system performance. We also point to a ticket where the OpenVPN issue is explained and log messages exist to diagnose this particular problem.

I can't spend time on this personally beyond a certain amount of community help as I am neither paid in something to do with OPNsense nor willing to cut short my off-work family time. I'm sorry.

Maybe someone else can. There is also a commercial offering and it is usually pretty good. But I feel that is not the point of this live debugging session anymore.


Cheers,
Franco

[mimugmail]

Can you Open a Bug Report in GitHub regarding Allowed Networks? I fix this after vacation