On 05/31/2011 07:42 PM, Sam Varshavchik wrote:
Tom Horsley writes:
On Tue, 31 May 2011 16:26:21 -0400 Sam Varshavchik wrote:
Before I begin a long and painful adventure in pulling apart with
what's
happening with systemd/initscripts, anyone has any clues where I
should look?
Check /etc/adjtime, it should say LOCAL, not UTC. You can also run hwclock --localtime to resync the hardware clock to local time.
It's says LOCAL, and the hwclock is most certainly resynced. Besides, at eash reboot hwclock-save.service should assure that the bios clock gets synced. But, at the next boot, it's broken again.
I just went through this with a Windows dual boot system, and I could swear I didn't say clock uses UTC when I installed, but maybe I did out of reflex.
I upgraded from F14. There were no issues before the upgrade.
I thought that it was odd that after I upgraded to F15, applied the updates, and rebooted a few times, the clock was off by twelve hours. I just shrugged it off as a gremlin, and resynced from my ntp server.
But when I noticed that the clock immediately jumped to being four hours later after a reboot, the alarm bell went off.
Unfortunately, now that we have this systemd infrastructure, it's not as easy as sticking a bunch of "date"s in the various rc scripts, to see what's happening to the clock when the system boots, and where it runs askew.
A brief Google search shows that I'm not the only one:
http://fedoraforum.org/forum/showthread.php?p=1475721#post1475721
… and, we're in Bugzilla. Looks like a bug to me.
Fixed? systemd may be the unintentional culprit. systemctl status ntpd.service