Thanks for this information. See below.
On 09/02/2015 04:42 AM, Miroslav Lichvar wrote:
On Tue, Sep 01, 2015 at 03:36:40PM -0400, Robert Moskowitz wrote:
Anyway. If I boot up without a network connection, the time stays at
Jan 1 1970.
I then set the time with the 'date' command and did:
touch /var/lib/systemd/clock
chown systemd-timesync:systemd-timesync /var/lib/systemd/clock
And rebooted. Time is once again at Jan 1 1970. I checked:
# ls -ls /var/lib/systemd/
total 16
4 drwxr-xr-x. 2 root root 4096 Aug 23 13:39 catalog
0 -rw-r--r--. 1 systemd-timesync systemd-timesync 0 Sep 1 15:27 clock
4 drwxr-xr-x. 2 root root 4096 May 21 15:06 coredump
4 -rw-------. 1 root root 512 Jan 1 1970
random-seed
4 drwxr-xr-x. 2 root root 4096 Aug 23 13:58 timers
So clock has a good timestamp. Maybe the wrong chown names? Or
something still yet?
Is systemd-timesyncd actually running? You need to disable chronyd
first as the two services conflict with each other.
OH?!? I am running F22 with Xfce, last updated Aug 23. So whatever is
in there wrt systemd-timesyscd and chronyd is what was in the image
orginally plus whatever happened when I did the date/time config in the
graphic configurator as part of the install (I go in and change my city
from NYC to Detroit). A quick check shows:
#systemctl status systemd-timesyncd.service
● systemd-timesyncd.service - Network Time Synchronization
Loaded: loaded (/usr/lib/systemd/system/systemd-timesyncd.service;
disabled; vendor preset: disabled)
Active: inactive (dead)
Docs: man:systemd-timesyncd.service(8)
# systemctl -l status chronyd
● chronyd.service - NTP client/server
Loaded: loaded (/usr/lib/systemd/system/chronyd.service; enabled;
vendor preset: enabled)
Active: active (running) since Thu 1970-01-01 00:00:21 EST; 2min 57s ago
Process: 482 ExecStartPost=/usr/libexec/chrony-helper update-daemon
(code=exited, status=0/SUCCESS)
Process: 451 ExecStart=/usr/sbin/chronyd $OPTIONS (code=exited,
status=0/SUCCESS)
Main PID: 459 (chronyd)
CGroup: /system.slice/chronyd.service
└─459 /usr/sbin/chronyd
Jan 01 00:00:19 vega.htt-consult.com systemd[1]: Starting NTP
client/server...
Jan 01 00:00:19 vega.htt-consult.com chronyd[459]: chronyd version 2.1.1
starting (+CMDMON +NTP +REFCLOCK +RTC +PRIVDROP +DEBUG +ASYNCDNS +IPV6
+SECHASH)
Jan 01 00:00:20 vega.htt-consult.com chronyd[459]: Frequency -2.808 +/-
0.045 ppm read from /var/lib/chrony/drift
Jan 01 00:00:21 vega.htt-consult.com systemd[1]: Started NTP client/server.
So it seems both are present, but only chronyd is running.
But you don't need to switch to timesyncd to initialize the system
time from a file on boot, chronyd will do that too with the -s option
when RTC is missing.
Create /etc/sysconfig/chronyd if it doesn't exist and put OPTIONS="-s"
there. It seems you do have an RTC, but you don't want to use it
as it keeps bad time, so you will also need to tell chronyd to ignore
it, for example by adding "rtcdevice /dev/nonexist" to
/etc/chrony.conf. On boot chronyd should now set the system time to
the modification time of /var/lib/chrony/drift.
I made these changes. In /etc/chrony.conf I now have:
# Enable kernel synchronization of the real-time clock (RTC).
#rtcsync
rtcdevice /dev/nonexist
I commented out the rtcsync line, given you told me to add the nonexist
line. Hope that is correct.
Further, looking at this file I see that chronyd can act as an ntp
server for other systems on the LAN? Important to know, no more ntpd then.
Perhaps these settings should be standard in our armv7 builds? Or an
easy option to set them. A check box for 'no rtc' on that configurator?
_______________________________________________
arm mailing list
arm@xxxxxxxxxxxxxxxxxxxxxxx
https://admin.fedoraproject.org/mailman/listinfo/arm
