F12-alpha: superblock always in future

Todd Denniston Todd.Denniston at tsb.cranrdte.navy.mil
Tue Sep 1 23:44:15 UTC 2009


sean darcy wrote, On 09/01/2009 06:16 PM:
> OK, let's say UTC is correct. I don't care. What do I change ( remember 
> it won't boot) from a Live CD that will allow it to boot?
> 
> Also (venting) why does it keep retreating 4 hours into the past? That 
> is, when I boot it tells me last mount was 4 hours in future ( and 
> remember, this is first boot from install).
> So I fsck it, changing the timestamp. Now on reboot, it's still 4 hours 
> in the future, because system clock has gone back another 4 hours. This 
> is a feature?
> 

This (retreating time) does seem rather odd...
Mind trying an experiment?
[assuming an FS tune2fs can read]
At no time while running the experiment should you attempt to SET the time.

Boot into BIOS, record (on paper) the shown time.
Boot into the Live CD, run date, record (on paper) the shown time.
	tune2fs -l /dev/devicewithproblem |grep -B1 time
	[record times]
Boot into BIOS, record the shown time.
Boot into the Live CD, run date, record the shown time.
	tune2fs -l /dev/devicewithproblem |grep -B1 time
	[record times]
Boot into BIOS, record the shown time.
Boot into the installed system, if you can run date, record the shown time.
	head -1 /etc/adjtime # record and report this line.
	#even if you can't get to the command line,
	#reboot and get the next BIOS run.
Boot into BIOS, record the shown time.

report from experiment:
times against each of the events above.
Live CD name & version.

I expect to see
First BIOS < First Live CD < second BIOS < second Live CD < third BIOS
I expect (from you indications) that tune2fs should show a time greater than any of the BIOS and the 
Live CDs.

if First BIOS < First Live CD but First BIOS > second BIOS, then I would expect that the halt script 
is mucking about with hwclock when it should not be (in my opinion the halt script should not mess 
with the hwclock unless it can prove that the system clock is truly synchronized with a good time 
source).

I am wondering if a hwclock --adjust is being called in the installed system, and the drift rate is 
huge.

BTW the file that hwclock uses to determine UTC/local time when reading/writing the hardware clock 
is /etc/adjtime.
  From the hwclock man page about /etc/adjtime:
        Line  3:  "UTC" or "LOCAL".  Tells whether the Hardware Clock is set to
        Coordinated Universal Time or local time.


-- 
Todd Denniston
Crane Division, Naval Surface Warfare Center (NSWC Crane)
Harnessing the Power of Technology for the Warfighter




More information about the fedora-test-list mailing list