Topic: [SOLVED] SME 8.x: Issue with NTPD or misconfiguration?

[Michail Pappas]

Running a stable 8.1 box for some 3 years now. Yesterday I saw there was an update for ntp and after installing I tried to check status with "peers" and "rl" from within the ntpq shell, but I receive only a timeout:

Code: [Select]

ntpq> peers
localhost: timed out, nothing received
***Request timed out

Additionally, checking the logs up to yesterday, I could not locate any log lines of the form "ntpd[xxx]: synchronized" form.

Furthermore, yesterday I got confused and restarted the ntpd service with /etc/init.d/ntpd restart, resulting:

Code: [Select]

# /etc/init.d/ntpd restart
Shutting down ntpd:                                        [  OK  ]
ntpd: Synchronizing with time server:                      [FAILED]
Starting ntpd:                                             [  OK  ]

config show ntpd:

Code: [Select]

# config show ntpd
ntpd=service
    NTPServer=X.Y.Z.W
    SyncToHWClockSupported=yes
    status=enabled

It's not an issue of the actual NTP server I'm getting time from:

Code: [Select]

# sv stop ntpd
ok: down: ntpd: 1s

# ntpdate -q X.Y.Z.W
server X.Y.Z.W, stratum 4, offset -0.000024, delay 0.02608
30 Dec 08:18:29 ntpdate[2461]: adjust time server X.Y.Z.W offset -0.000024 sec

I did find this though, while grep'ing for ntp in /var/log:

Code: [Select]

secure.20141222011202:Dec 29 08:42:24 myserver useradd[14722]: failed adding user `ntp', data deleted

Seems like a script went wrong, or did I do something bad?

[Stefano]

Michail: if something isn't working out-of-the-box as expected, bugzilla is right there.. thank you

[guest22]

Seems like a script went wrong, or did I do something bad?

As per Stefano's advise, please add it to the bugtracker, including what you see in /var/log/messages and /var/log/ntpd

[mmccarn]

You're running into bug 8517

ntpd (the service) is running, but is configured by default to not allow *any* client access.

To allow client (ntpq, ntpstat) access to your ntpd from the sme server command line:

Code: [Select]

mkdir -p /etc/e-smith/templates-custom/etc/ntp.conf
cd /etc/e-smith/templates-custom/etc/ntp.conf
echo 'restrict 127.0.0.1' > 41allowLocalhost
expand-template /etc/ntp.conf
sv t ntpd


[guest22]

You're running into bug 8517

next to that, I believe there is something else going on. Hence the request for the logs. It seems that it is possible there is a race condition for binding to 0.0.0.0 between ntpd and ntpdate, where ntpdate won't release it.
But we can discuss this once a there is evidence for it so we can raise a bug against it.

[Michail Pappas]

I believe mmcarn nailed it. Only one question though: if this is a feature and not a bug, why does starting the service with sv start ntpd throw an error (see original post)?

[guest22]

I believe mmcarn nailed it. Only one question though: if this is a feature and not a bug, why does starting the service with sv start ntpd throw an error (see original post)?

I suspect of what I said in my previous comment.

[Daniel B.]

Original post talks about restarting ntpd with /etc/init.d/ntpd restart. This is expected to fail, ntpd is monitored and must be restarted using sv t /services/ntpd

[Michail Pappas]

After a reboot I could not reproduce any errors, nothing bad I could see in /var/log/messages or else. Consider this thread solved please.