Re: [chrony-users] What's changed in chrony-2.2?

[Peter Humphrey <peter@xxxxxxxxxxxxxxx>]

On Monday 23 November 2015 08:15:14 Miroslav Lichvar wrote:
> On Sun, Nov 22, 2015 at 04:30:55PM +0000, Peter Humphrey wrote:
> > My two main machines are an i5 workstation and an Atom mini-server, both
> > running Gentoo stable. This week chrony was updated from 2.1.1 to 2.2,
> > and now it won't start on the Atom though it seems fine on the i5.
> 
> What exactly does it do? Crash on segfault, a fatal error, an
> assertion error, or something else? As Bill says, there might be
> something in the syslog.

Sorry, I should have been clearer.

# /etc/init.d/chronyd start
 * Starting chronyd ...
 * start-stop-daemon: caught an interrupt
 * start-stop-daemon: /usr/sbin/chronyd died
 * Failed to start chronyd                                          [ !! ]
 * ERROR: chronyd failed to start

But a chronyd process is actually running:

20169 ?        S      0:00 /usr/sbin/chronyd -f /etc/chrony/chrony.conf

# grep chronyd /var/log/messages
--->8
Nov 23 09:45:50 serv chronyd[20096]: chronyd exiting
Nov 23 09:46:02 serv chronyd[20169]: chronyd version 2.2 starting (+CMDMON 
+NTP +REFCLOCK +RTC -PRIVDROP -SCFILTER -SECHASH +ASYNCDNS +IPV6 -
DEBUG)
Nov 23 09:46:02 serv chronyd[20169]: Frequency -28.615 +/- 0.000 ppm read 
from /var/lib/chrony/drift
Nov 23 09:46:02 serv /etc/init.d/chronyd[20166]: start-stop-daemon: caught 
an interrupt
Nov 23 09:46:02 serv /etc/init.d/chronyd[20166]: start-stop-daemon: 
/usr/sbin/chronyd died
Nov 23 09:46:02 serv /etc/init.d/chronyd[20149]: ERROR: chronyd failed to 
start
Nov 23 09:46:07 serv chronyd[20169]: Selected source 212.23.8.6
Nov 23 09:48:18 serv chronyd[20169]: Selected source 176.126.242.239

> How do your chrony.conf and chronyd command line look like?

This is chrony.conf:

pool uk.pool.ntp.org iburst
server ntp0.zen.co.uk iburst
server ntp1.zen.co.uk iburst
driftfile /var/lib/chrony/drift
makestep 1.0 3
allow 192.168/16
rtcsync
port 0

It's the same on both machines, and so of course is /etc/init.d/chronyd, 
whose start code is here (checkconfig and setxtrarg are internal functions 
which do what they say):

start() {
        checkconfig || return $?
        setxtrarg

        [ -n "${PIDFILE}" ] || PIDFILE=/run/chronyd.pid

        ebegin "Starting chronyd"
        start-stop-daemon --start --background --quiet \
                --exec /usr/sbin/chronyd \
                --pidfile "${PIDFILE}" \
                -- -f "${CFGFILE}" ${ARGS}
        eend $? "Failed to start chronyd"
}

> > Has something changed in chrony's kernel config requirements?
> 
> The most significant change between 2.1.1 and 2.2 is that chronyd
> creates a Unix domain socket in /var/run/chrony. I see in your kernel
> config it's a module, maybe it's not loaded?

Do you mean RTC_CMOS is a module? It's the same on both machines, and a 
socket is indeed created:

# ls -l /var/run/chrony
total 0
srwxr-xr-x 1 root root 0 Nov 23 09:46 chronyd.sock
# lsmod | grep rtc
rtc_cmos                7264  0

> But even if it wasn't I think it would just send an error message to
> syslog and continue without the command socket.

It may be doing something like that here too, but with the socket.

> In Fedora the SELinux rules had to be updated to allow chronyd to
> create the socket, but that didn't prevent it from running (with
> reduced functionality).

-- 
Rgds
Peter


-- 
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx 
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx 
with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.